Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Claim 1,
Claim 1 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 1 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“identifying one or more statistics associated with each of multiple processing regions within the image, each processing region representing a portion of the image”
“generating a probability of each of the processing regions containing at least one object of interest based on the statistics associated with the processing regions”
“allocating multiple processing windows to one or more of the processing regions based on the probabilities, the processing windows smaller than the processing regions”
“performing object detection within the allocated processing windows”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are insignificant extra-solution activity (See MPEP 2106.05(g)).
The limitations:
“obtaining an image of a scene”
As drafted, are additional elements that amount to no more than insignificant extra-solution activity. See MPEP 2106.05(g).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “insignificant extra-solution activity”. Specifically, the obtaining limitation recites the well-understood, routine, and conventional activity of receiving and transmitting data over a network. MPEP 2106.05(d)(II); OIP Techs., Inc v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network). Insignificant extra-solution activity cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 2,
Claim 2 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 2 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“processing the statistics associated with the processing regions”
“convert the statistics associated with the processing regions into the probabilities”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“using a machine learning model”
“the machine learning model trained to…”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 3,
Claim 3 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 3 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: See corresponding analysis of claim 2.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein the machine learning model comprises a support vector machine (SVM) classifier”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 4,
Claim 4 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 4 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“convert distances from a hyperplane classification boundary associated with the SVM classifier into probabilities along a continuous scale”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein the machine learning model further comprises an activation function configured to”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 5,
Claim 5 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 5 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“convert the statistics associated with the processing regions into the probabilities using a labeled training dataset, the labeled training dataset comprising training images that are known to contain objects, training images that are known to not contain objects, and labels indicating which of the training images contain and do not contain objects”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“training the machine learning model to…”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 6,
Claim 6 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 6 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: See corresponding analysis of claim 5.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein training the machine learning model comprises using stochastic gradient descent to minimize hinge loss across the labeled training dataset while using a ridge regularization of parameters of the machine learning model”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 7,
Claim 7 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 7 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“determining a number of processing windows to allocate to each of the processing regions based on a ratio involving (i) a specified statistic associated with the processing region and (ii) a sum of the specified statistic across all of the processing regions”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: See corresponding analysis of claim 1.
Step 2B Analysis: See corresponding analysis of claim 1.
Regarding Claim 8,
Claim 8 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 8 is directed to an apparatus comprising at least one memory and at least one processor, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“identify one or more statistics associated with each of multiple processing regions within the image, each processing region representing a portion of the image”
“generate a probability of each of the processing regions containing at least one object of interest based on the statistics associated with the processing regions”
“allocate multiple processing windows to one or more of the processing regions based on the probabilities, the processing windows smaller than the processing regions”
“perform object detection within the allocated processing windows”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“An apparatus comprising: at least one memory configured to store an image of a scene; and at least one processing device configured to…”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 9,
Claim 9 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 9 is directed to an apparatus comprising at least one memory and at least one processor, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“process the statistics associated with the processing regions”
“convert the statistics associated with the processing regions into the probabilities”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“the at least one processing device is configured to…”
“using a machine learning model”
“the machine learning model trained to…”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 10,
Claim 10 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 10 is directed to an apparatus comprising at least one memory and at least one processor, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: See corresponding analysis of claim 9.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein the machine learning model comprises a support vector machine (SVM) classifier”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 11,
Claim 11 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 11 is directed to an apparatus comprising at least one memory and at least one processor, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“convert distances from a hyperplane classification boundary associated with the SVM classifier into probabilities along a continuous scale”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein the machine learning model further comprises an activation function configured to”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 12,
Claim 12 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 12 is directed to an apparatus comprising at least one memory and at least one processor, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“convert the statistics associated with the processing regions into the probabilities using a labeled training dataset, the labeled training dataset comprising training images that are known to contain objects, training images that are known to not contain objects, and labels indicating which of the training images contain and do not contain objects”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein the at least one processing device is further configured to train the machine learning model to”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 13,
Claim 13 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 13 is directed to an apparatus comprising at least one memory and at least one processor, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: See corresponding analysis of claim 12.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein, to train the machine learning model, the at least one processing device is configured to use stochastic gradient descent to minimize hinge loss across the labeled training dataset and use a ridge regularization of parameters of the machine learning model”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 14,
Claim 14 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 14 is directed to an apparatus comprising at least one memory and at least one processor, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“determine a number of processing windows to allocate to each of the processing regions based on a ratio involving (i) a specified statistic associated with the processing region and (ii) a sum of the specified statistic across all of the processing regions”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein, to allocate the processing windows to the one or more processing regions, the at least one processing device is configured to…”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 15,
Claim 15 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 15 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“generate probabilities that processing regions within captured images contain at least one object, each processing region representing a portion of the corresponding captured image”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)).
The limitations:
“training a machine learning model to…”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
The limitations:
“obtaining a labeled training dataset, the labeled training dataset comprising training images that are known to contain objects, training images that are known to not contain objects, and labels indicating which of the training images contain and do not contain objects”
As drafted, are additional elements that amount to no more than insignificant extra-solution activity. See MPEP 2106.05(g).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply” or “insignificant extra-solution activity”. Specifically, the obtaining limitation recites the well-understood, routine, and conventional activity of receiving and transmitting data over a network. MPEP 2106.05(d)(II); OIP Techs., Inc v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network). Mere instructions to apply an exception and insignificant extra-solution activity cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 16,
Claim 16 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 16 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“convert statistics associated with the processing regions into the probabilities”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein the machine learning model is trained to…”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 17,
Claim 17 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 17 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: See corresponding analysis of claim 15.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein training the machine learning model comprises using stochastic gradient descent to minimize hinge loss across the labeled training dataset and using a ridge regularization of parameters of the machine learning model”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 18,
Claim 18 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 18 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: See corresponding analysis of claim 15.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein the machine learning model comprises a support vector machine (SVM) classifier”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 19,
Claim 19 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 19 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“convert distances from a hyperplane classification boundary associated with the SVM classifier into probabilities along a continuous scale”
As drafted, under their broadest reasonable interpretations, cover mental processes, i.e., concepts performed in the human mind (including an observation, evaluation, judgement, opinion). The above limitations in the context of this claim correspond to mental processes, e.g., evaluation and judgement with assistance of pen and paper.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recited additional elements that are mere instructions to apply an exception (See MPEP 2106.05(f)).
The limitations:
“wherein the machine learning model further comprises an activation function configured to”
As drafted, are additional elements that amount to no more than mere instructions to apply an exception for the abstract ideas. See MPEP 2106.05(f).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract ideas into a practical application, all of the additional elements are “mere instructions to apply”. Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 20,
Claim 20 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 20 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: See corresponding analysis of claim 15.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are insignificant extra-solution activity (See MPEP 2106.05(g)).
The limitations:
“deploying the trained machine learning model to a platform for use in performing object detection”
As drafted, are additional elements that amount to no more than insignificant extra-solution activity. See MPEP 2106.05(g).
Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “insignificant extra-solution activity”. Specifically, the deploying limitation recites the well-understood, routine, and conventional activity of receiving and transmitting data over a network. MPEP 2106.05(d)(II); OIP Techs., Inc v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network). Insignificant extra-solution activity cannot provide an inventive concept. The claim is not patent eligible.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1 and 8 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Liu et al. (U.S. Patent Publication No. 2021/0303862) (“Liu”).
Regarding claim 1, Liu teaches a method comprising: obtaining an image of a scene (Liu [0003] “According to the method, image data representing a scene is obtained and sound distribution information related to the scene is obtained.”; [0053] “Specifically, the information obtaining module 410 is configured to obtain image data 402 representing a scene. Object detection is to be performed on the image data 402. The image data 402 may include a separate digital image or consist of a sequence of consecutive digital image (i.e., a video). The image data 402 may capture or represent a part or a whole scene. For example, the image data 402 may represent a physical environment. In some embodiments, the image data 402 may be captured by one or more image capturing devices deployed in the scene, such as a camera, a panoramic camera, and/or the like.” Liu provides obtaining images of a scene with information obtaining module 410.); identifying one or more statistics associated with each of multiple processing regions within the image (Liu [0050] “The sound distribution information is helpful in the object detection as it mimics human-like object detection. In real life, a real, patrolman, for example, would always hear a sound produced in a scene first and then search and focus on the source of the sound to check what happened. Through the solution for object detection proposed herein, the sound distribution information can be used to guide the object detection to focus on image regions that are of higher probabilities of representing a target object(s).”; [0056] “In some embodiments, the sound distribution information 404 may be represented in form of a heat map, as illustrated in the example of FIG. 4. In some embodiments, the sound distribution information 404 may include sound energy distribution information to indicate sound energy levels (or power levels) distributed across the scene… The elements of the first heat map may be corresponding to pixels in an image with values indicating the sound energy levels.” Liu provides identifying sound distribution information that correspond to pixels in an image of a scene, corresponding to identifying one or more statistics associated with each of multiple processing regions within the image.), each processing region representing a portion of the image (Liu [0050] “Through the solution for object detection proposed herein, the sound distribution information can be used to guide the object detection to focus on image regions that are of higher probabilities of representing a target object(s).” Liu provides region based image analysis, corresponding to each processing region representing a portion of the image.); generating a probability of each of the processing regions containing at least one object of interest based on the statistics associated with the processing regions (Liu [0067] “The strategy determination module 420 may then determine the detection strategy 422 in which the object detection is enabled to be focused on or pay more attention to the region of interest. For example, finer or more complex object detection may be applied to the region of interest as compared with other image regions. This is because the region of interest may have a higher probability of representing or partially representing a target object to be detected. As such, instead of making equal efforts to detect a target object across the whole image or the whole video, the sound distribution information can help locate an important region(s) or frame(s) in a video on which the object detection should focus.” Liu provides strategy determination module 420 which generates probabilities for regions in an image based on associated sound distribution information, corresponding to generating a probability of each of the processing regions containing at least one object of interest based on the statistics associated with the processing regions, wherein the sound distribution information corresponds to the statistics.); allocating multiple processing windows to one or more of the processing regions based on the probabilities (Liu [0050] “Through the solution for object detection proposed herein, the sound distribution information can be used to guide the object detection to focus on image regions that are of higher probabilities of representing a target object(s). Such sound-aware object detection can improve accuracy of the object detection, possibly improve the computation efficiency and reduce the computation power by allocating more computation resources to perform detection on the image regions of higher probabilities of representing a target object(s).” Liu provides allocating more computation resources to perform object detection on image regions of higher probabilities of representing a target object based on sound distribution information, corresponding to allocating multiple processing windows to one or more of the processing regions based on the probabilities.), the processing windows smaller than the processing regions (Liu [0075] “For example, as shown in FIG. 6B, the strategy determination module 420 may determine smaller sizes of bounding boxes 602 for the region of interest 512 in the digital image of the image data 402. As such, the region of interest may be searched with finger granularity as compared with the remaining region.” Liu provides bounding boxes that are smaller than a region of interest in an image, corresponding to the processing windows are smaller than the processing regions.); and performing object detection within the allocated processing windows (Liu [0050] “Through the solution for object detection proposed herein, the sound distribution information can be used to guide the object detection to focus on image regions that are of higher probabilities of representing a target object(s). Such sound-aware object detection can improve accuracy of the object detection, possibly improve the computation efficiency and reduce the computation power by allocating more computation resources to perform detection on the image regions of higher probabilities of representing a target object(s).”; [0081] “At block 710, the system 400 obtains image data representing a scene and sound distribution information related to the scene. At block 720, the system 400 determines a detection strategy to be applied in object detection based on the sound distribution information. At block 730, the system 400 performs the object detection on the image data by applying the detection strategy.” Liu provides performing object detection within image regions of higher probability of representing a target object, wherein more computational resources are allocated to the regions of higher probability of a target object using a detection strategy, corresponding to performing object detection within the allocated processing windows, wherein the “processing windows” correspond to a fixed number of computational resources allocated to specific portions of an image for object detection, as discussed in paragraph [0024] of the specification.).
Regarding claim 8, it is the apparatus embodiment of claim 1 with similar limitations to claim 1 and is rejected using the same reasoning found above in the rejection of claim 1. Further, Liu teaches an apparatus comprising: at least one memory configured to store an image of a scene (Liu [0053] “Specifically, the information obtaining module 410 is configured to obtain image data 402 representing a scene.”; [0052] “The system 400 includes an information obtaining module 410, a strategy determination module 420, and an object detection module 430. The system 400 may be implemented by computer system/server 12 of FIG. 1 and the modules 410 to 430 in the system 400 may be implemented in software, hardware, middleware, and/or any combination thereof.”; [0044] “Examples of hardware components include: mainframes 61; RISC (Reduced Instruction Set Computer) architecture based servers 62; servers 63; blade servers 64; storage devices 65; and networks and networking components 66. In some embodiments, software components include network application server software 67 and database software 68.” Liu provides obtaining images of a scene including storage devices corresponding to at least one memory configured to store an image of a scene.); and at least one processing device (Liu [0091] “The present invention may be a system, a method, and/or a computer program product at any possible technical detail level of integration. The computer program product may include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present invention.” Liu provides at least one processing device.).
Claims 15-16 and 18 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Kozitsky et al. (U.S. Patent Publication No. 2017/0262723) (“Kozitsky”).
Regarding claim 15, Kozitsky teaches a method comprising: obtaining a labeled training dataset, the labeled training dataset comprising training images that are known to contain objects, training images that are known to not contain objects, and labels indicating which of the training images contain and do not contain objects (Kozitsky [0037] “The strong classifier is trained to discriminate between readable and unreadable license plate images. The readable image set includes all regions that contain a license plate, which can he recognized by a human, and in turn capable of being successfully processed with an ALPR engine. The unreadable set includes all regions for which a human cannot recognize the license plate code and/or state. This set would include cases where the license plate is not present, is partially occluded, is too dark, too bright, or mangled, etc. Our goal is to automatically identify and exclude the un-readable (non-revenue) imagery from human review.”; [0040] “Note that “SVM” (Support Vector Machine) is a machine learning supervised learning model with associated learning algorithms that together can analyze data for classification and/or regression analysis. Given a set of training examples, each marked for belonging to one of two categories, an SVM training algorithm can build a model that assigns new examples into one category or the other, making it a non-probabilistic binary linear classifier.” Kozitsky provides a obtaining a training dataset for training an SVM classifier, which includes images in which a license plate is present and images in which a license plate is not present, including corresponding labels, respectively corresponding to the labeled training dataset comprising training images that are known to contain objects, training images that are known to not contain objects, and labels indicating which of the training images contain and do not contain objects.); and training a machine learning model to generate probabilities that processing regions within captured images contain at least one object, each processing region representing a portion of the corresponding captured image (Kozitsky [0037] “The strong classifier is trained to discriminate between readable and unreadable license plate images.”; [0038] “In an example embodiment, each ROI image can be resized to 224×224 before passing to the CNN defined and trained and extracting the 4096 features of the layer before softmax. These features can be used as input to a linear SVM trained to differentiate between ROI's with plates and those without. Platt's method can be used to convert the SVM score to a posterior probability and this probability can then be employed as the confidence output of the second stage.”; [0040] “Note that “SVM” (Support Vector Machine) is a machine learning supervised learning model with associated learning algorithms that together can analyze data for classification and/or regression analysis. Given a set of training examples, each marked for belonging to one of two categories, an SVM training algorithm can build a model that assigns new examples into one category or the other, making it a non-probabilistic binary linear classifier.” Kozitsky provides training a support vector machine (SVM) to generate probabilities using Platt’s method, which converts SVM output to probabilities, to detect license plates in obtained image, corresponding to training a machine learning model to generate probabilities that processing regions within captured images contain at least one object, each processing region representing a portion of the corresponding captured image.).
Regarding claim 16, Kozitsky teaches the method of claim 15, wherein the machine learning model is trained to convert statistics associated with the processing regions into the probabilities (Kozitsky [0037] “The strong classifier is trained to discriminate between readable and unreadable license plate images.”; [0038] “These features can be used as input to a linear SVM trained to differentiate between ROI's with plates and those without. Platt's method can be used to convert the SVM score to a posterior probability and this probability can then be employed as the confidence output of the second stage.” Kozitsky provides an SVM trained to convert image features into region of interest probabilities associated with license plates, corresponding to the machine learning model is trained to convert statistics associated with the processing regions into the probabilities.).
Regarding claim 18, Kozitsky teaches the method of claim 15, wherein the machine learning model comprises a support vector machine (SVM) classifier (Kozitsky [0039] “FIG. 5 illustrates a block diagram depicting a strong classifier 80 applied to each ROI (Region of Interest) in a sample image 82, in accordance with an example embodiment. FIG. 5 depicts the process flow of the second stage for each ROI. As shown at block 84, the image 82 is resized and then subject to a CNN 86. Output from the CNN is sent to a linear SVM (Support Vector Machine) module 88. Output from the linear SVM module 88 can be processed by a posterior probability module 90 and the resulting confidence 92 generated.”; [0040] “Note that “SVM” (Support Vector Machine) is a machine learning supervised learning model with associated learning algorithms that together can analyze data for classification and/or regression analysis.” Kozitsky provides a machine learning model comprises an SVM classifier.).
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 2-4 and 9-11 are rejected under 35 U.S.C. 103 as being unpatentable over Liu et al. (U.S. Patent Publication No. 2021/0303862) (“Liu”) in view of Larlus-Larrondo et al. (U.S. Patent Publication No. 2014/0037198) (“Larlus-Larrondo”).
Regarding claim 2, Liu teaches the method of Claim 1 as discussed above in the rejection of claim 1, but fails to explicitly teach wherein generating the probabilities comprises: processing the statistics associated with the processing regions using a machine learning model, the machine learning model trained to convert the statistics associated with the processing regions into the probabilities.
However, Larlus-Larrondo teaches wherein generating the probabilities comprises: processing the statistics associated with the processing regions using a machine learning model, the machine learning model trained to convert the statistics associated with the processing regions into the probabilities (Larlus-Larrondo [0061] “At S134 a score or a probability at the pixel level is provided for each class, using the classification outputs (e.g., scores) for all regions, triplets, and triplet parents that contain a given pixel, to compute the probability for that pixel. In the case of classifier models based on support vector machines (SVMs), for example, a sigmoid function can be used to convert the classifier SVM outputs into probabilities… Then, the probabilities are aggregated, e.g., averaged, over all regions and triplets containing a pixel to obtain the probability for that pixel.” Larlus-Larrondo provides a support vector machine (SVM) corresponding to a machine learning model, wherein a sigmoid function converts the SVM classifier outputs into probabilities associated with pixels in an image, corresponding to processing the statistics associated with the processing regions using a machine learning model, the machine learning model trained to convert the statistics associated with the processing regions into the probabilities.).
Liu and Larlus-Larrondo are both considered to be analogous to the claimed invention because they are in the same field of artificial intelligence and more specifically applied to object detection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Liu with the above teachings of Larlus-Larrondo. Doing so would allow for improved image segmentation results (Larlus-Larrondo [0023] “A system and method for semantic image segmentation are disclosed that combine a description of the image in a hierarchy of regions and suitable learning mechanisms to improve the results of segmentation.”).
Regarding claim 3, Liu in view of Larlus-Larrondo teaches the method of Claim 2 as discussed above in the rejection of claim 2, wherein the machine learning model comprises a support vector machine (SVM) classifier (Larlus-Larrondo [0061] “At S134 a score or a probability at the pixel level is provided for each class, using the classification outputs (e.g., scores) for all regions, triplets, and triplet parents that contain a given pixel, to compute the probability for that pixel. In the case of classifier models based on support vector machines (SVMs), for example, a sigmoid function can be used to convert the classifier SVM outputs into probabilities… Then, the probabilities are aggregated, e.g., averaged, over all regions and triplets containing a pixel to obtain the probability for that pixel.” Larlus-Larrondo provides an SVM classifier.).
Liu and Larlus-Larrondo are both considered to be analogous to the claimed invention because they are in the same field of artificial intelligence and more specifically applied to object detection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Liu with the above teachings of Larlus-Larrondo. Doing so would allow for improved image segmentation results (Larlus-Larrondo [0023] “A system and method for semantic image segmentation are disclosed that combine a description of the image in a hierarchy of regions and suitable learning mechanisms to improve the results of segmentation.”).
Regarding claim 4, Liu in view of Larlus-Larrondo teaches the method of claim 3 as discussed above in the rejection of claim 3, wherein the machine learning model further comprises an activation function configured to convert distances from a hyperplane classification boundary associated with the SVM classifier into probabilities along a continuous scale (Larlus-Larrondo [0061] “At S134 a score or a probability at the pixel level is provided for each class, using the classification outputs (e.g., scores) for all regions, triplets, and triplet parents that contain a given pixel, to compute the probability for that pixel. In the case of classifier models based on support vector machines (SVMs), for example, a sigmoid function can be used to convert the classifier SVM outputs into probabilities… Then, the probabilities are aggregated, e.g., averaged, over all regions and triplets containing a pixel to obtain the probability for that pixel.” Larlus-Larrondo provides a sigmoid function, corresponding to an activation function, which is configured to convert SVM output, corresponding to distances from a hyperplane classification boundary associated with the SVM classifier, into probabilities using the sigmoid function, which produces a continuous scale, since the “SVM outputs” are obtained by calculating distances from a hyperplane classification boundary, which are then converted into probabilities using a sigmoid function, which outputs continuous values (i.e., any value between (0, 1), therefore corresponding to an activation function configured to convert distances from a hyperplane classification boundary associated with the SVM classifier into probabilities along a continuous scale.).
Liu and Larlus-Larrondo are both considered to be analogous to the claimed invention because they are in the same field of artificial intelligence and more specifically applied to object detection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Liu with the above teachings of Larlus-Larrondo. Doing so would allow for improved image segmentation results (Larlus-Larrondo [0023] “A system and method for semantic image segmentation are disclosed that combine a description of the image in a hierarchy of regions and suitable learning mechanisms to improve the results of segmentation.”).
Regarding claim 9, the rejection of claim 8 is incorporated herein. Further, the limitations in this claim are taught by Liu in view of Larlus-Larrondo for the same reasons disclosed above in the rejection of claim 2.
Regarding claim 10, the rejection of claim 9 is incorporated herein. Further, the limitations in this claim are taught by Liu in view of Larlus-Larrondo for the same reasons disclosed above in the rejection of claim 3.
Regarding claim 11, the rejection of claim 10 is incorporated herein. Further, the limitations in this claim are taught by Liu in view of Larlus-Larrondo for the same reasons disclosed above in the rejection of claim 4.
Claims 5 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over Liu et al. (U.S. Patent Publication No. 2021/0303862) (“Liu”) in view of Larlus-Larrondo et al. (U.S. Patent Publication No. 2014/0037198) (“Larlus-Larrondo”) in further view of Kozitsky et al. (U.S. Patent Publication No. 2017/0262723) (“Kozitsky”).
Regarding claim 5, Liu in view of Larlus-Larrondo teaches the method of Claim 2, as discussed above in the rejection of claim 2, but fails to explicitly teach further comprising: training the machine learning model to convert the statistics associated with the processing regions into the probabilities using a labeled training dataset, the labeled training dataset comprising training images that are known to contain objects, training images that are known to not contain objects, and labels indicating which of the training images contain and do not contain objects.
However, Kozitsky teaches training the machine learning model to convert the statistics associated with the processing regions into the probabilities using a labeled training dataset (Kozitsky [0038] “In an example embodiment, each ROI image can be resized to 224×224 before passing to the CNN defined and trained and extracting the 4096 features of the layer before softmax. These features can be used as input to a linear SVM trained to differentiate between ROI's with plates and those without. Platt's method can be used to convert the SVM score to a posterior probability and this probability can then be employed as the confidence output of the second stage.”; [0040] “Note that “SVM” (Support Vector Machine) is a machine learning supervised learning model with associated learning algorithms that together can analyze data for classification and/or regression analysis. Given a set of training examples, each marked for belonging to one of two categories, an SVM training algorithm can build a model that assigns new examples into one category or the other, making it a non-probabilistic binary linear classifier.” Kozitsky provides training an SVM using a set of training examples to perform Platt scaling (Platt’s method) for image analysis, which converts SVM output to probabilities, which is then used for image-processing and video-based detection, corresponding to training the machine learning model to convert the statistics associated with the processing regions into the probabilities using a labeled training dataset.), the labeled training dataset comprising training images that are known to contain objects, training images that are known to not contain objects, and labels indicating which of the training images contain and do not contain objects (Kozitsky [0037] “The strong classifier is trained to discriminate between readable and unreadable license plate images. The readable image set includes all regions that contain a license plate, which can he recognized by a human, and in turn capable of being successfully processed with an ALPR engine. The unreadable set includes all regions for which a human cannot recognize the license plate code and/or state. This set would include cases where the license plate is not present, is partially occluded, is too dark, too bright, or mangled, etc. Our goal is to automatically identify and exclude the un-readable (non-revenue) imagery from human review.”; [0040] “Note that “SVM” (Support Vector Machine) is a machine learning supervised learning model with associated learning algorithms that together can analyze data for classification and/or regression analysis. Given a set of training examples, each marked for belonging to one of two categories, an SVM training algorithm can build a model that assigns new examples into one category or the other, making it a non-probabilistic binary linear classifier.” Kozitsky provides a training dataset for training an SVM classifier, which includes images in which a license plate is present and images in which a license plate is not present, including corresponding labels, respectively corresponding to the labeled training dataset comprising training images that are known to contain objects, training images that are known to not contain objects, and labels indicating which of the training images contain and do not contain objects.).
Liu, Larlus-Larrondo, and Kozitsky are all considered to be analogous to the claimed invention because they are in the same field of artificial intelligence and more specifically applied to object detection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Liu in view of Larlus-Larrondo with the above teachings of Kozitsky. Doing so would allow for a trained SVM that can analyze data for classification and/or regression analysis (Kozitsky [0040] “Note that “SVM” (Support Vector Machine) is a machine learning supervised learning model with associated learning algorithms that together can analyze data for classification and/or regression analysis.”).
Regarding claim 12, the rejection of claim 9 is incorporated herein. Further, the limitations in this claim are taught by Liu in view of Larlus-Larrondo in further view of Kozitsky for the same reasons disclosed above in the rejection of claim 5.
Claims 6 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over Liu et al. (U.S. Patent Publication No. 2021/0303862) (“Liu”) in view of Larlus-Larrondo et al. (U.S. Patent Publication No. 2014/0037198) (“Larlus-Larrondo”) in further view of Kozitsky et al. (U.S. Patent Publication No. 2017/0262723) (“Kozitsky”) and Engelcke et al. (U.S. Patent Publication No. 2020/0019794) (“Engelcke”).
Regarding claim 6, Liu in view of Larlus-Larrondo in further view of Kozitsky teaches the method of claim 5, as discussed above in the rejection of claim 5, but fails to teach wherein training the machine learning model comprises using stochastic gradient descent to minimize hinge loss across the labeled training dataset while using a ridge regularization of parameters of the machine learning model.
However, Engelcke teaches wherein training the machine learning model comprises using stochastic gradient descent to minimize hinge loss across the labeled training dataset while using a ridge regularization of parameters of the machine learning model (Engelcke [0135] “Each of the three class specific networks 136a-c is a binary classifier and it is therefore appropriate to use a linear hinge loss for training due to its maximum margin property. In the embodiment being described, the hinge loss, L2 weight decay and an L1 sparsity penalty are used to train the networks with stochastic gradient descent. Both the L2 weight decay as well as the L1 sparsity penalty serve as regularisers.”; [0137] “Given an output detection score x.sub.0 and a class label y∈{−1, 1} distinguishing between positive and negative samples, the hinge loss is formulated as: Eq. (14).” Engelcke provides training machine learning models including using stochastic gradient descent, a corresponding hinge loss, and L2 regularization (ridge regularization) and labeled training data, wherein L2 regularization is also known as ridge regularization, which penalizes loss functions with the sum of squares.)
Liu, Larlus-Larrondo, Kozitsky, and Engelcke are all considered to be analogous to the claimed invention because they are in the same field of artificial intelligence and more specifically applied to object detection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Liu in view of Larlus-Larrondo in further view of Kozitsky with the above teachings of Engelcke. Doing so would allow for a maximum margin property during training (Engelcke [0135] “Each of the three class specific networks 136a-c is a binary classifier and it is therefore appropriate to use a linear hinge loss for training due to its maximum margin property”).
Regarding claim 13, the rejection of claim 12 is incorporated herein. Further, the limitations in this claim are taught by Liu in view of Larlus-Larrondo in further view of Kozitsky and Engelcke for the same reasons disclosed above in the rejection of claim 6.
Claims 7 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Liu et al. (U.S. Patent Publication No. 2021/0303862) (“Liu”) in view of Rhee et al. (U.S. Patent Publication No. 2022/0012588) (“Rhee”).
Regarding claim 7, Liu teaches the method of claim 1 as discussed above in the rejection of claim 1, but fails to teach wherein allocating the processing windows to the one or more processing regions comprises: determining a number of processing windows to allocate to each of the processing regions based on a ratio involving (i) a specified statistic associated with the processing region and (ii) a sum of the specified statistic across all of the processing regions.
However, Rhee teaches wherein allocating the processing windows to the one or more processing regions comprises: determining a number of processing windows to allocate to each of the processing regions based on a ratio involving (i) a specified statistic associated with the processing region and (ii) a sum of the specified statistic across all of the processing regions (Rhee [0056] “The time-series data stream 200 may be one of the most natural input types based on a time law. The data stream 200 may be a flow of information, such as, for example, text classification, questions and answers, language instruction, translation, object detection, subtitle caption, video representation, and the like. The data stream 200 is described herein as an image in a vision field, and label information is described herein as including a label value indicating a class corresponding to a type of an object included in the image.”; [0075] “The reservoir management device may determine an ideal target partition ratio p.sub.i of each class with respect to i∈[u], as represented by Equation 1 below”; [0076] “The reservoir management device may determine a target label distribution 519 that indicates the target partition ratio p.sub.i for each class based on an occurrence frequency n.sub.i observed for each class in the data stream 590 and an allocation exponent ρ. The target label distribution 519 may be a set of target partition ratios, for example, [p.sub.1, . . . p.sub.i, . . . p.sub.u]. As represented by Equation 1 above, the reservoir management device may calculate a target partition ratio p.sub.i for an i-th class to be a ratio of a value obtained by raising an occurrence frequency n.sub.i observed for the i-th class to the allocation exponent ρ-th power (n.sub.i raised to the ρ-th power) to a sum of values obtained by raising the occurrence frequency n.sub.j observed for each class to the allocation exponent ρ-th power (n.sub.j raised to the ρ-th power). Here, j denotes an index indicating a set of classes observed up to the current time point in the data stream 590, and may be an integer greater than or equal to 1 and less than or equal to u.” Rhee provides Equation (1), which is used for determining resource allocation in object detection, and includes a ratio, as shown in Equation (1), which includes an occurrence frequency for each class as the numerator and a sum of values obtained by raising the occurrence frequency n.sub.j observed for each class to the allocation exponent ρ-th power as the denominator, corresponding to the allocation comprises a ratio involving (i) a specified statistic associated with the processing region and (ii) a sum of the specified statistic across all of the processing regions.).
Liu and Rhee are both considered to be analogous to the claimed invention because they are in the same field of artificial intelligence and more specifically applied to object detection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Liu with the above teachings of Rhee. Doing so would allow for an allocation of sufficient memory space for a given task (Rhee [0058] “The reservoir management device may allocate a sufficient memory space even to moderate and minor labels, rather than major labels, in a limited and fixed memory space.”).
Regarding claim 14, the rejection of claim 8 is incorporated herein. Further, the limitations in this claim are taught by Liu in view of Rhee for the same reasons disclosed above in the rejection of claim 7.
Claim 17 is rejected under 35 U.S.C. 103 as being unpatentable over Kozitsky et al. (U.S. Patent Publication No. 2017/0262723) (“Kozitsky”) in view of Engelcke et al. (U.S. Patent Publication No. 2020/0019794) (“Engelcke”).
Regarding claim 17, Kozitsky teaches the method of claim 15 as discussed above in the rejection of claim 15, but fails to teach wherein training the machine learning model comprises using stochastic gradient descent to minimize hinge loss across the labeled training dataset and using a ridge regularization of parameters of the machine learning model.
However, Engelcke teaches wherein training the machine learning model comprises using stochastic gradient descent to minimize hinge loss across the labeled training dataset and using a ridge regularization of parameters of the machine learning model (Engelcke [0135] “Each of the three class specific networks 136a-c is a binary classifier and it is therefore appropriate to use a linear hinge loss for training due to its maximum margin property. In the embodiment being described, the hinge loss, L2 weight decay and an L1 sparsity penalty are used to train the networks with stochastic gradient descent. Both the L2 weight decay as well as the L1 sparsity penalty serve as regularisers.”; [0137] “Given an output detection score x.sub.0 and a class label y∈{−1, 1} distinguishing between positive and negative samples, the hinge loss is formulated as: Eq. (14).” Engelcke provides training machine learning models including using stochastic gradient descent, a corresponding hinge loss, and L2 regularization (ridge regularization) and labeled training data, wherein L2 regularization is also known as ridge regularization, which penalizes loss functions with the sum of squares.).
Kozitsky, and Engelcke are both considered to be analogous to the claimed invention because they are in the same field of artificial intelligence and more specifically applied to object detection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Kozitsky with the above teachings of Engelcke. Doing so would allow for a maximum margin property during training (Engelcke [0135] “Each of the three class specific networks 136a-c is a binary classifier and it is therefore appropriate to use a linear hinge loss for training due to its maximum margin property”).
Claim 19 is rejected under 35 U.S.C. 103 as being unpatentable over Kozitsky et al. (U.S. Patent Publication No. 2017/0262723) (“Kozitsky”) in view of Larlus-Larrondo et al. (U.S. Patent Publication No. 2014/0037198) (“Larlus-Larrondo”).
Regarding claim 19, Kozitsky teaches the method of claim 18 as discussed above in the rejection of claim 18, but fails to explicitly teach wherein the machine learning model further comprises an activation function configured to convert distances from a hyperplane classification boundary associated with the SVM classifier into probabilities along a continuous scale.
However, Larlus-Larrondo teaches wherein the machine learning model further comprises an activation function configured to convert distances from a hyperplane classification boundary associated with the SVM classifier into probabilities along a continuous scale (Larlus-Larrondo [0061] “At S134 a score or a probability at the pixel level is provided for each class, using the classification outputs (e.g., scores) for all regions, triplets, and triplet parents that contain a given pixel, to compute the probability for that pixel. In the case of classifier models based on support vector machines (SVMs), for example, a sigmoid function can be used to convert the classifier SVM outputs into probabilities… Then, the probabilities are aggregated, e.g., averaged, over all regions and triplets containing a pixel to obtain the probability for that pixel.” Larlus-Larrondo provides a sigmoid function, corresponding to an activation function, which is configured to convert SVM output, corresponding to distances from a hyperplane classification boundary associated with the SVM classifier, into probabilities using the sigmoid function, which produces a continuous scale, since the “SVM outputs” are obtained by calculating distances from a hyperplane classification boundary, which are then converted into probabilities using a sigmoid function, which outputs continuous values (i.e., any value between (0, 1), therefore corresponding to an activation function configured to convert distances from a hyperplane classification boundary associated with the SVM classifier into probabilities along a continuous scale.).
Kozitsky and Larlus-Larrondo are both considered to be analogous to the claimed invention because they are in the same field of artificial intelligence and more specifically applied to object detection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Kozitsky with the above teachings of Larlus-Larrondo. Doing so would allow for improved image segmentation results (Larlus-Larrondo [0023] “A system and method for semantic image segmentation are disclosed that combine a description of the image in a hierarchy of regions and suitable learning mechanisms to improve the results of segmentation.”).
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Kozitsky et al. (U.S. Patent Publication No. 2017/0262723) (“Kozitsky”) in view of Liu et al. (U.S. Patent Publication No. 2021/0303862) (“Liu”).
Regarding claim 20, Kozitsky teaches the method of claim 15 as discussed above in the rejection of claim 15, but fails to explicitly teach further comprising: deploying the trained machine learning model to a platform for use in performing object detection.
However, Liu teaches deploying the trained machine learning model to a platform for use in performing object detection (Liu [0025] “Platform as a Service (PaaS): the capability provided to the consumer is to deploy onto the cloud infrastructure consumer-created or acquired applications created using programming languages and tools supported by the provider.”; [0053] “Specifically, the information obtaining module 410 is configured to obtain image data 402 representing a scene. Object detection is to be performed on the image data 402. The image data 402 may include a separate digital image or consist of a sequence of consecutive digital image (i.e., a video). The image data 402 may capture or represent a part or a whole scene. For example, the image data 402 may represent a physical environment. In some embodiments, the image data 402 may be captured by one or more image capturing devices deployed in the scene, such as a camera, a panoramic camera, and/or the like.” Liu provides deploying a trained machine learning to a platform for performing object detection.).
Kozitsky and Liu are both considered to be analogous to the claimed invention because they are in the same field of artificial intelligence and more specifically applied to object detection. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Kozitsky with the above teachings of Liu. Doing so would allow for a consumer to have control over the deployed applications and possibly application hosting environment configurations (Liu [0025] “The consumer does not manage or control the underlying cloud infrastructure including networks, servers, operating systems, or storage, but has control over the deployed applications and possibly application hosting environment configurations.”).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KURT NICHOLAS PRESSLY whose telephone number is (703)756-4639. The examiner can normally be reached M-F 8-4.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Kamran Afshar can be reached at (571) 272-7796. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/KURT NICHOLAS PRESSLY/Examiner, Art Unit 2125
/KAMRAN AFSHAR/Supervisory Patent Examiner, Art Unit 2125