DETAILED ACTION
Notice of Pre-AIA or AIA Status
1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
2. This is in response to the applicant response filed on 04/24/2026. In the applicant’s response, claims 1, 3, 4, 10, 12, 14, 18, and 19 were amended; claims 2, 11, 13, 20, and 21 were cancelled. Accordingly, claims 1, 3-10, 12, and 14-19 are pending and being examined. Claims 1, 10, 12, 14, 18, and 19 are independent form.
3. The rejections of the claims under 35 USC 101 have been withdrawn in view of applicant’s amendment.
Claim Rejections - 35 USC § 112
4. The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
5. Claims 1, 3-10, 12, and 14-19 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre- AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
5-1. Regarding independent claim 1, the claim recites: “generating focus information for learning generation, wherein information based on each of the acquired plurality of learning images is input to the focal position estimation model during training, a calculation is performed according to the focal position estimation model to acquire information indicating a focal position when in focus according to a position in each of the plurality of learning images, and focus information for learning indicating a focal position when in focus according to a position in an image used for machine learning training is generated from the acquired information and the in-focus position information for each of the plurality of learning images” in lines 14-22. It appears that what the applicant means is that the focal position estimation model via “a calculation” acquires both the “information indicating a focal position” and the “focus information for learning” for each of the acquired plurality of learning images. However, it is not clear to which information the acquired information refers. In other words, it is not clear whether the acquired information refers to the information indicating a focal position, the focus information for learning indicating a focal position, or the both. The claim(s) do/does not define the metes and bound of the claimed invention with a reasonable degree of precision and particularity, and thus is/are rejected under 35 U.S.C. 112(b).
5-2. Regarding independent claims 10, 12, 14, 18, and 19, each of them faces the same issue set forth in the rejection of independent claim 1, and thus, is rejected as being indefinite under 35 U.S.C. 112(b).
5-3. The remaining claims are dependent from claims 1, 10, 12, 14, 18, or 19, respectively, therefore, are rejected as being indefinite under 35 U.S.C. 112(b).
5-4. Although the claims are indefinite, the examiner is interpreting and examining the claims by the examiner’s best understandings for the purpose of examination.
Claim Rejections - 35 USC § 103
6. 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 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.
7. 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 of this title, 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.
8. Claims 1, 3-10, 12, and 14-19 are rejected under 35 U.S.C. 103 as being unpatentable over Sabato et al (US 2022/0383525, hereinafter “Sabato”) in view of Pinkarp et al (“Deep learning for single-shot autofocus microscopy”, 2019, hereinafter “Pinkarp”).
Regarding claim 1, Sabato discloses a focal position estimation method for estimating a focal position when in focus corresponding to an estimation target image (“the computer-implementable method for extracting depth information from a plurality of images taken by a camera at different focus positions” through the CNN 720 shown by fig.1; see para.9 and fig.5), the method comprising:
acquiring an estimation target image (see 803 of fig.5 and para.229: “processing, by the machine learning algorithm [i.e., the convolutional neural network (CNN) disclosed by fig.2a], a captured image whose image features have not yet been extracted, said captured image representing a currently processed image, e.g. input image 101.”); and
estimating a focal position when in focus at which an imaging target captured in an image is in focus corresponding to the estimation target image and according to a position in the estimation target image, from the acquired estimation target image, by using a focal position estimation model that is generated through machine learning training and that receives information based on an image as its input and outputs information indicating the focal position when in focus according to a position in the image (see 807 of fig.3 and para.233: “Generating a two-dimensional depth map using the focus positions specified in the predetermined focus schedule and the at least one generated multi-dimensional tensor”. See para.92: wherein “[t]he step of generating a two-dimensional depth map [...] may further comprise, generating, by the machine learning algorithm, at least one multi-dimensional focus probability map [...]” In other words, the machine learning algorithm may estimate one multi-dimensional focus position probability map for the input image using the focus positions specified in the predetermined focus schedule and the at least one generated multi-dimensional tensor.);
wherein the focal position estimation model is generated by: acquiring a plurality of learning images of the same imaging target at different focal positions, each of which is associated with a focal position, and in-focus position information indicating focal positions when in focus for the plurality of learning images (see the training phase disclosed by fig.6 and para.234—para.244: which comprises capturing a sequence of images focused at different focus positions according to a focus schedule for each of a plurality of different scenes from the real physical world (see para.10: “capturing a sequence of images of a scene with a camera at different focus positions according to a predetermined focus schedule that specifies a chronological sequence of focus positions of the camera”), obtaining a sequence of focus probability maps, one for each image after a predetermined number of captured images have been processed, remapping the obtained focus probability maps to real distances using the focus positions from the known focus schedule, calculating the loss function between the estimated/predicted depth maps are with respect to the expected known ground truth focus depth maps, and minimizing the loss function and to determine the best model parameters of the machine learning algorithm/convolutional neural network from the training images.).
Although the CNN model of Sabato does not trained by “generating focus information for learning generation, wherein information based on each of the acquired plurality of learning images is input to the focal position estimation model during training, a calculation is performed according to the focal position estimation model to acquire information indicating a focal position when in focus according to a position in each of the plurality of learning images, and focus information for learning indicating a focal position when in focus according to a position in an image used for machine learning training is generated from the acquired information and the in-focus position information for each of the plurality of learning images; and learning, wherein machine learning training for generating the focal position estimation model is performed by using the information based on each of the acquired plurality of learning images and the generated focus information for learning corresponding to each of the plurality of learning images generated” as recited in the claim. However, in the same field of endeavor, Pinkarp teaches a FCFNN model, wherein the FCFNN model is generated by:
acquiring a plurality of learning images of the same imaging target at different focal positions, each of which is associated with a focal position, and in-focus position information indicating focal positions when in focus for the plurality of learning images (see fig.1 (a)—Training data);
generating focus information for learning generation, wherein information based on each of the acquired plurality of learning images is input to the focal position estimation model during training (see “Ground truth focal position” in the left of fig.1 (a), wherein the ground truth focal position is generated based on the incoherent images of the training data), a calculation is performed according to the focal position estimation model to acquire information indicating a focal position when in focus according to a position in each of the plurality of learning images, and focus information for learning indicating a focal position when in focus according to a position in an image used for machine learning training is generated from the acquired information and the in-focus position information for each of the plurality of learning images; and learning, wherein machine learning training for generating the focal position estimation model is performed by using the information based on each of the acquired plurality of learning images and the generated focus information for learning corresponding to each of the plurality of learning images generated (wherein the “neural network” shown by the right of fig.1 (a) is trained/learned based on the ground truth focal position (i.e., the in-focus position information) acquired by the model and each coherent image of the training data via calculations.).
It would have been obvious to one of ordinary skill in the art before the effective filling date of the claimed invention was made to incorporate the teachings of Pinkarp into the teachings of Sabato and train the neural network model based an output result of the neural network model and in-focus position information as taught by Pinkarp. Suggestion or motivation for doing so would have been to provide “a fast, robust autofocusing method” as taught by Pinkarp, see the Title and Abstract. Therefore, the combination of Sabato and Pinkarp teaches or suggests all the limitations recited in claim 1, and the claim is unpatentable over Sabato in view of Pinkarp.
Regarding claim 3, 15, the combination of Sabato and Pinkarp discloses, wherein the focal position estimation model is generated by calculating one focal position when in focus, which is common to the plurality of learning images, according to a position in each learning image from the focal position when in focus according to a position in each of the plurality of learning images indicated by the information acquired by using the focal position estimation model during training and generating the focus information for learning from the one focal position when in focus for each of the plurality of learning images (see 801-807 of fig.5: the convolutional neural network (CNN) “[g]enerat[es] a two-dimensional depth map [which is “common” to each of the sequence of images captured at step 801 of fig.5] using the focus positions specified in the predetermined focus schedule and the at least one generated multi-dimensional tensor”; see para.226-para.233).
Regarding claim 4, 16, the combination of Sabato and Pinkarp discloses, wherein, a feature quantity of the estimation target image is acquired from the acquired estimation target image by using a feature quantity output model that receives information based on an image as its input and outputs a feature quantity of the image input to the focal position estimation model, and the focal position when in focus corresponding to the estimation target image and according to a position in the estimation target image is estimated from the feature quantity by using the focal position estimation model (e.g., Sabato, see, e.g.,para.235: “a training sample comprising a plurality/a sequence 600 of captured images focused at different focus positions according to a focus schedule 620 for a plurality of different scenes from the real physical world can be processed according to the steps described previously to obtain a sequence 640 of focus probability maps, one for each image after a predetermined number of captured images have been processed.”), and
the feature quantity output model is generated by generating two different feature quantity learning images, which are associated with focal positions and correspond to the plurality of learning images, based on information indicating the focal position when in focus according to a position in each of the plurality of learning images, which is acquired by using the focal position estimation model during the training, comparing feature quantities of the two feature quantity learning images with each other according to focal positions associated with the two feature quantity learning images with a combination of the two feature quantity learning images as one unit, and performing machine learning training based on a result of the comparison (e.g., Sabato, see para.24-para.242: The loss function 660 is a measure of how different the estimated/predicted depth maps are with respect to the expected known ground truth depth maps.” “The training of the machine learning algorithm 630 comprising a convolutional neural network is run until the loss function has reached a desired/specified minimum and the optimal model parameters of the convolutional neural network have been determined.” “The minimization of the loss function may be achieved by optimization techniques such as using a gradient descent algorithm.”).
Regarding claim 5, 17, the combination of Sabato and Pinkarp discloses, wherein the feature quantity output model is generated by performing the machine learning training so that a difference between the feature quantities of the two feature quantity learning images becomes smaller when the two feature quantity learning images are related to the same focal position and the difference between the feature quantities of the two feature quantity learning images becomes larger when the two feature quantity learning images are related to different focal positions (ibid.).
Regarding claim 6, the combination of Sabato and Pinkarp discloses the focal position estimation method according to claim 1, wherein, an inclination of an imaging target captured in the estimation target image is estimated from the estimated focal position when in focus according to a position in the estimation target image (e.g., Sabato, see para.237: “The scenes captured in the sequence 600 of images of the training sample can be static or dynamic, i.e. there can be movement between images, e.g. due to movement of objects or subjects in the scene and/or due to movement of the camera, e.g. vibrations due to the camera being held in the hand of a user or due to the camera changing its position.”).
Regarding claim 7, the combination of Sabato and Pinkarp discloses the focal position estimation method according to claim 1, wherein, a focal position when imaging an imaging target captured in the estimation target image is controlled based on the estimated focal position when in focus according to a position in the estimation target image (e.g., Sabato, see para.235: “a training sample comprising a plurality/a sequence 600 of captured images focused at different focus positions according to a focus schedule 620 for a plurality of different scenes from the real physical world can be processed according to the steps described previously to obtain a sequence 640 of focus probability maps, one for each image after a predetermined number of captured images have been processed.”).
Regarding claim 8, the combination of Sabato and Pinkarp discloses the focal position estimation according to claim 1, wherein, information indicating an in-focus state according to a position in the estimation target image is output based on the estimated focal position when in focus according to a position in the estimation target image (e.g., Sabato, see para.233: “Generating a two-dimensional depth map [i.e., the output of the CNN] using the focus positions specified in the predetermined focus schedule and the at least one generated multi-dimensional tensor”).
Regarding claim 9, the combination of Sabato and Pinkarp discloses the focal position estimation method according to claim 1, wherein, a plurality of estimation target images of the same imaging target at different focal positions are acquired (e.g., Sabato, see 801 of fig.5 and para.227: “Capturing, 801, a sequence of images of a scene [i.e., the same target] with a camera at different focus positions according to a predetermined focus schedule that specifies a chronological sequence of focus positions of the camera”), and a focal position when in focus according to a position in the estimation target image is estimated from at least one estimation target image among the acquired plurality of estimation target images, and one image is generated from the plurality of estimation target images based on the estimated focal position (e.g., Sabato, see 807 of fig.3 and para.233: “Generating a two-dimensional depth map using the focus positions specified in the predetermined focus schedule and the at least one generated multi-dimensional tensor”. See para.92: wherein “[t]he step of generating a two-dimensional depth map [...] may further comprise, generating, by the machine learning algorithm, at least one multi-dimensional focus probability map [...]”.).
Regarding claims 10, 12, each of them is an inherent variation of claim 1, thus it is interpreted and rejected for the reasons set forth in the rejection of claim 1.
Regarding claim 14, 18, 19, the essential features of each of them are recited by claim 1 or by the combination of claim 1 and claim 2, thus they are interpreted and rejected for the reasons set forth in the rejections of claim 1 and claim 2.
Response to Arguments
9. Applicant's arguments have been considered but are moot in view of the new ground(s) of rejection.
Conclusion
10. Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the date of this final action.
11. Any inquiry concerning this communication or earlier communications from the examiner should be directed to RUIPING LI whose telephone number is (571)270-3376. The examiner can normally be reached 8:30am--5:30pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, HENOK SHIFERAW can be reached on (571)272-4637. 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; 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.
/RUIPING LI/Primary Examiner, Ph.D., Art Unit 2676
Prosecution Insights