DETAILED ACTION
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 .
Response to Arguments
Applicant’s arguments, see Remarks page 8, filed 7/21/2026, with respect to the Objections of Claims 5 & 16 have been fully considered and are persuasive. The Objections of Claims 5 & 16 have been withdrawn.
Applicant’s arguments, see Remarks page 8, filed 7/21/2026, with respect to the Rejections of Claims 6-7 under 35 U.S.C. 112(b) have been fully considered and are persuasive. The Rejections of Claims 6-7 have been withdrawn.
Applicant’s arguments, see Remarks pages 8-10, filed 7/21/2026, with respect to the rejections of claim(s) 1, 14, and 17 under 35 U.S.C. 102(a)(1) have been fully considered and are moot in view of the new grounds of rejection (detailed in the rejections below) necessitated by Applicant’s amendment to the claim(s).
Applicant’s arguments, see Remarks pages 11-12, filed 7/21/2026, with respect to the rejections of claim(s) 6 & 7 under 35 U.S.C. 103 have been fully considered and are moot in view of the new grounds of rejection (detailed in the rejections below) necessitated by Applicant’s amendment to the claim(s).
Claim Objections
Claim 7 is objected to because of the following informalities:
Regarding claim 7, the limitation “determine an incremental object mask as the difference between the combined object mask and one or more existing object masks.” Should be corrected to “determine an incremental object mask as a difference between the combined object mask and one or more existing object masks.”
Appropriate correction is required.
Claim Rejections - 35 USC § 102
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.
Claim(s) 1 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Wei et al. (Fusion of an Ensemble of Augmented Image Detectors for Robust Object Detection) hereinafter referenced as Wei.
Regarding claim 1, Wei discloses: A system comprising: one or more processors to execute a processing pipeline to detect an object based at least on iteratively transforming and processing input data that includes at least one representation of the object (Wei: Abstract: “Herein, a new computational intelligence fusion approach based on the dynamic analysis of agreement among object detection outputs is proposed. Furthermore, we propose an online versus just in training image augmentation strategy…The approach is demonstrated in the context of cone, pedestrian and box detection for Advanced Driver Assistance Systems (ADAS) applications.”;
Algorithm 1: “
PNG
media_image1.png
5
508
media_image1.png
Greyscale
”), and aggregating results of each of at least two iterations to generate a combined object detection result (Wei: 1. Introduction: “we hypothesize that presenting a deep network with multiple augmented images during testing and fusing the results could result in a more robust detection system. Herein, we develop a system utilizing image augmentation combined with Axis-Aligned Bounding Box Fuzzy Integral (AABBFI)-based fusion to enhance the detection results. The fusion method applied in this paper originates from the field of computational intelligence. This method analyses the agreement among object detection outputs and fuses them dynamically.”), wherein for a subsequent iteration following an initial iteration, the one or more processors are further to update at least one value of one or more data transformation parameters relative to a value used in an iteration immediately preceding the subsequent iteration the initial iteration and to generate new transformed data by applying a data transformation to original input data based on the at least one updated value (Wei: Algorithm 1: “
PNG
media_image1.png
5
508
media_image1.png
Greyscale
”;
Section: 4.2.1. Augmentation Methods: “In the experiments performed, the Python Imaging Library (PIL) is utilized [40] to obtain augmented images. In PIL brightness and contrast enhancement classes, a factor is used for the change of brightness and contrast. When this factor is one, it gives the original image. Based on the experimental evaluation, in ADAS examples, factor values are chosen to be 1, 0.25, 0.5, 1.5, 2.0 and 2.5 for brightness and contrast. To add Gaussian noise, the noise variance is chosen to be 0.001, 0.003 and 0.005. This is based on qualitative image assessment, since these noise levels do not drastically alter the appearance of the input image. Other augmentation methods, including edge enhancement, global histogram equalization and Gaussian blurring (radius = 2), are predefined image operations in PIL.”).
Claim Rejections - 35 USC § 103
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.
Claim(s) 1-2, 4-5, 8-10, and 13-19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Hu et al. (US 2022/0028087 A1) hereinafter referenced as Hu, in view of Wei.
Regarding claim 1, Hu discloses: A system comprising: one or more processors to execute a processing pipeline to detect an object based at least on iteratively transforming and processing input data that includes at least one representation of the object (Hu: Abstract; 0045: “the terminal device 10 a obtains a target image including a target object, the target object being a to-be-segmented object… the backend server 10 d obtains an original mask and a trained image segmentation model. Downsampling is performed on the original mask based on a first unit model in the image segmentation model to obtain a downsampled mask. Region convolution feature information of the target image is extracted based on a second unit model in the image segmentation model and the downsampled mask, and the original mask is updated according to the region convolution feature information, the original mask being updated continuously and iteratively. When it is detected that an updated original mask satisfies an error convergence condition, the backend server 10 d determines an image region of the target object in the target image according to the updated original mask.”), and aggregating results of each of at least two iterations to generate a combined object detection result (Hu: 0061: “ for each unit mask in the mask 20 g , it may be determined that a value of the unit mask is greater than a mask threshold. In response to determining that the value of the unit mask is greater than the mask threshold, the value of the unit mark is adjusted to 1; and in response to determining that the value of the unit mask is not greater than the mask threshold, the value of the unit mark is adjusted to 0. It may be learned that a binary mask is obtained after the adjustment, and the binary mask is used as a new original mask.”;
0063: “In some implementations, the mask threshold may be a semi-dynamic value, which may be determined by the image segmentation model based on the image 20 a and/or the original mask 20 b.”; Wherein the resulting new original mask is based on a combination of a previous original mask and a newly determined mask.).
Hu does not disclose expressly: wherein for a subsequent iteration following an initial iteration, the one or more processors are further to update at least one value of one or more data transformation parameters relative to a value used in an iteration immediately preceding the subsequent iteration and to generate new transformed data by applying a data transformation to original input data based on the at least one updated value.
Wei discloses: a method for performing image-based object detection by fusing the detection results of a plurality of object detection algorithms (Wei: Abstract: “A significant challenge in object detection is accurate identification of an object’s position in image space, whereas one algorithm with one set of parameters is usually not enough, and the fusion of multiple algorithms and/or parameters can lead to more robust results. Herein, a new computational intelligence fusion approach based on the dynamic analysis of agreement among object detection outputs is proposed. Furthermore, we propose an online versus just in training image augmentation strategy.”). Wherein the method further comprises an iterative image augmentation method, wherein for a subsequent iteration following an initial iteration, one or more processors are configured to update at least one value of one or more data transformation parameters relative to a value used in an iteration immediately preceding the subsequent iteration and to generate new transformed data by applying a data transformation to original input data based on the at least one updated value (Wei: Algorithm 1: “
PNG
media_image1.png
5
508
media_image1.png
Greyscale
”;
Section: 4.2.1. Augmentation Methods: “In the experiments performed, the Python Imaging Library (PIL) is utilized [40] to obtain augmented images. In PIL brightness and contrast enhancement classes, a factor is used for the change of brightness and contrast. When this factor is one, it gives the original image. Based on the experimental evaluation, in ADAS examples, factor values are chosen to be 1, 0.25, 0.5, 1.5, 2.0 and 2.5 for brightness and contrast. To add Gaussian noise, the noise variance is chosen to be 0.001, 0.003 and 0.005. This is based on qualitative image assessment, since these noise levels do not drastically alter the appearance of the input image. Other augmentation methods, including edge enhancement, global histogram equalization and Gaussian blurring (radius = 2), are predefined image operations in PIL.”).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to implement the algorithms for image augmentation prior to performing object detection as taught by Wei in order to generate augmented target images for the iterative system for generating object masks disclosed by Hu. The suggestion/motivation for doing so would have been “For augmentation of each input, the goal is to produce a range of inputs. The “optimal” augmentation cannot be determined algorithmically, and it varies depending on the type of objects detected, the image background, etc. Instead, a range of images of varying quality is generated and presented to the detector. The augmentation methods used herein include changing brightness, contrast, edge enhancement, global histogram equalization, Gaussian blurring and adding independent and identically distributed (IID) Gaussian noise to simulate different scenarios…we choose to focus on basic operations, which we have already shown lead to success.” (Wei: 3.2 Augmentation). Further, one skilled in the art could have combined the elements as described above by known methods with no change in their respective functions, and the combination would have yielded nothing more than predictable results. Therefore, it would have been obvious to combine Hu with Wei to obtain the invention as specified in claim 1.
Regarding claim 2, Hu in view of Wei discloses: The system of claim 1, wherein for at least one iteration, the one or more processors are further to: perform object detection on the new transformed data according to one or more values of one or more object detection parameters to generate an object mask (Hu: 0056-0059: “ The terminal device 10 a inputs the original mask 20 b into the first unit model, and a pooling operation is performed on the original mask 20 b by using the pooling layer 1 in the first unit model…As shown in FIG. 2a, a mask 20 f is obtained after a pooling operation is performed by the pooling layer on the original mask 20 b with a size of 4×4, and a size of the mask 20 f is reduced to 2×2.
The terminal device 10 a inputs the image 20 a into the second unit model, and a convolution operation and a pooling operation are performed on the image 20 a by using the convolutional layer 1 and the pooling layer 1 in the second unit model to extract convolution feature information 20 d of the image 20 a . A size of a feature map of the convolution feature information 20 d is 2×2, and a quantity of channels is 4. A convolution operation and a pooling operation are performed on the convolution feature information 20 d by using the convolutional layer 2 and the pooling layer 2 in the second unit model and the mask 20 f after the pooling to extract deeper convolution feature information 20 e of the image 20 a . A size of a feature map of the deeper convolution feature information 20 e is also 2×2, and a quantity of channels is also 4, that is, a size of input data of the convolutional layer 2 and the pooling layer 2 is the same as a size of output data thereof…
A deconvolution operation is finally performed on convolution feature information extracted at the last time by using a transpose convolution layer, so that a size of a feature map after the deconvolution is the same as the size of the image 20 a . By using a fully connected layer, the feature information after the deconvolution is mapped into a mask 20 g , and a size of the mask 20 g is the same as the size of the image 20 a.”;
0064-0065: “So far, one update on the original mask 20 b is completed. Then the new original mask is inputted into the first unit model, and similarly, the image 20 a is inputted into the second unit model, and the original mask is updated again by using the foregoing manners, continuously and iteratively. When a difference quantity between an original mask before the update and an original mask after the update is less than a difference quantity threshold, the original mask updated at the last time is used as a target mask 20 g.”; Wherein the image, which is transformed as taught by Wei, and the iteratively updated original mask serve as object detection parameters.).
Regarding claim 4, Hu in view of Wei discloses: The system of claim 1, wherein the one or more processors are to perform the iteratively transforming and processing until a determination that a termination criteria has been satisfied (Hu: 0143: “the terminal device calculates an error between an original mask before the updating and an original mask after the updating, and detects whether the error is less than a preset error threshold. If the error is less than the preset error threshold, it indicates that the updated original mask satisfies the error convergence condition, and in this case, the terminal device may use the updated original mask as a target mask. If the error is not less than the preset error threshold, it indicates that the updated original mask does not satisfy the error convergence condition, and the original mask needs to be updated continuously and iteratively, until the updated original mask satisfies the error convergence condition.”), and wherein the aggregating of the results comprises merging object masks associated with two or more iterations to obtain a combined object mask (Hu: 0063: “In some implementations, the mask threshold may be a semi-dynamic value, which may be determined by the image segmentation model based on the image 20 a and/or the original mask 20 b.”).
Regarding claim 5, Hu in view of Wei discloses: The system of claim 2, wherein the one or more processors are further to: responsive to a determination that a termination criteria has not been satisfied, update at least one value of at least one of the one or more object detection parameters; and perform the object detection on the new transformed data according to the at least one updated value of the at least one of the one or more object detection parameters to obtain a new object mask (Hu:
0064-0065: “So far, one update on the original mask 20 b is completed. Then the new original mask is inputted into the first unit model, and similarly, the image 20 a is inputted into the second unit model, and the original mask is updated again by using the foregoing manners, continuously and iteratively. When a difference quantity between an original mask before the update and an original mask after the update is less than a difference quantity threshold, the original mask updated at the last time is used as a target mask 20 g.”; Wherein the image, which is transformed as taught by Wei, and the iteratively updated original mask serve as object detection parameters.).
Regarding claim 8, Hu in view of Wei discloses: The system of claim 1, wherein the iteratively transforming comprises applying a spatial filter to the input data (Wei: Figure 2; 4.2.1. Augmentation Methods: “In the experiments performed, the Python Imaging Library (PIL) is utilized [40] to obtain augmented images. In PIL brightness and contrast enhancement classes, a factor is used for the change of brightness and contrast. When this factor is one, it gives the original image. Based on the experimental evaluation, in ADAS examples, factor values are chosen to be 1, 0.25, 0.5, 1.5, 2.0 and 2.5 for brightness and contrast. To add Gaussian noise, the noise variance is chosen to be 0.001, 0.003 and 0.005. This is based on qualitative image assessment, since these noise levels do not drastically alter the appearance of the input image. Other augmentation methods, including edge enhancement, global histogram equalization and Gaussian blurring (radius = 2), are predefined image operations in PIL.”; Wherein augmentations, such as Gaussian blurring constitute spatial filters.).
Regarding claim 9, Hu in view of Wei discloses: The system of claim 1, wherein the one or more processors are further to: select a data transformation type corresponding to the iteratively transforming based at least on an object type corresponding to the object (Wei: 2.5. Object Detection Using Deep Learning in ADAS: “When using ADAS to help a driver in the process of driving, we first need the system to accurately detect objects near or far, big or small in various environmental conditions.”; 3.2 Augmentation: “For augmentation of each input, the goal is to produce a range of inputs. The “optimal” augmentation cannot be determined algorithmically, and it varies depending on the type of objects detected, the image background, etc. Instead, a range of images of varying quality is generated and presented to the detector. The augmentation methods used herein include changing brightness, contrast, edge enhancement, global histogram equalization, Gaussian blurring and adding independent and identically distributed (IID) Gaussian noise to simulate different scenarios…we choose to focus on basic operations, which we have already shown lead to success.”).
Regarding claim 10, Hu in view of Wei discloses: The system of claim 1, wherein the input data corresponds to a plurality of voxels and a result of the object detection is an object mask identifying at least a subset of the plurality of voxels corresponding to the object (Hu: 0075: “The target object may be a lesion object, and correspondingly, a target image including the lesion object may be a biological tissue image. The biological tissue image may be a computed tomography (CT) image, a magnetic resonance imaging (MRI) image, an endoscopic image, or the like, and the target image may be a two-dimensional image, a three-dimensional image, or the like.”;
0160-0161: “FIG. 7 is an architectural diagram of a model of an image processing method according to an embodiment of this application…A size of a target image is 192*168*128, and it indicates that the target image is a three-dimensional image. Correspondingly, an original mask is an all-0 three-dimensional image with a size of 192*168*128.”).
Regarding claim 13, Hu in view of Wei discloses: The system of claim 1, wherein the system is comprised in at least one of: a control system for an autonomous or semi-autonomous machine; a perception system for an autonomous or semi-autonomous machine; a system for performing simulation operations; a system for performing digital twin operations; a system for performing light transport simulation; a system for performing collaborative content creation for 3D assets; a system for presenting one or more of virtual reality content, augmented reality content, or mixed reality content; a system for hosting one or more real-time streaming applications; a system for performing deep learning operations; a system implemented using an edge device; a system implemented using a robot; a system for performing conversational AI operations; a system for generating synthetic data; a system incorporating one or more virtual machines (VMs); a system implemented at least partially in a data center; or a system implemented at least partially using cloud computing resources (Hu: 0044-0045: “FIG. 1 is a system architecture diagram of image processing according to an embodiment of this application. This application relates to a backend server 10 d and a terminal device cluster. For example, the terminal device cluster may include: a terminal device 10 a , a terminal device 10 b , . . . , and a terminal device 10 c. Using the terminal device 10 a as an example, the terminal device 10 a obtains a target image including a target object, the target object being a to-be-segmented object. The terminal device 10 a transmits the target image to the backend server 10 d , and after receiving the target image, the backend server 10 d obtains an original mask and a trained image segmentation model.”).
As per claim(s) 14, arguments made in rejecting claim(s) 1 are analogous.
As per claim(s) 15, arguments made in rejecting claim(s) 2 are analogous.
As per claim(s) 16, arguments made in rejecting claim(s) 4 are analogous.
As per claim(s) 17, arguments made in rejecting claim(s) 1 are analogous.
As per claim(s) 18, arguments made in rejecting claim(s) 2 are analogous.
Regarding claim 19, Hu in view of Wei discloses: The processor of claim 17, wherein a number of processing iterations of the plurality of processing iterations is determined based at least on a termination criteria (Hu: 0143: “the terminal device calculates an error between an original mask before the updating and an original mask after the updating, and detects whether the error is less than a preset error threshold. If the error is less than the preset error threshold, it indicates that the updated original mask satisfies the error convergence condition, and in this case, the terminal device may use the updated original mask as a target mask. If the error is not less than the preset error threshold, it indicates that the updated original mask does not satisfy the error convergence condition, and the original mask needs to be updated continuously and iteratively, until the updated original mask satisfies the error convergence condition.”).
Claim(s) 3, 11, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Hu in view of Wei, and further in view of Siriborvornratanakul et al. (Multiscale Visual Object Detection for Unsupervised Ubiquitous Projection Based on a Portable Projector-Camera System) hereinafter referenced as Siriborvornratanakul.
Regarding claim 3, Hu in view of Wei discloses: The system of claim 1.
Hu in view of Wei does not disclose expressly: wherein the one or more processors comprise parallel processing circuitry to accelerate application of the data transformation and performance of the object detection.
Siriborvornratanakul discloses: wherein the one or more processors comprise parallel processing circuitry to accelerate application of the data transformation (Siriborvornratanakul: Figure 6; Section III. D. Parallel implementation: “Considering recent growth in multicore processors and parallel programming languages, we changed the sequential implementation to the equivalent parallel implementation. The concept of our parallel implementation to the multiscale visual detection is illustrated in Fig. 6. The preprocessed image is distributed simultaneously to all scales. Instead of using the same smoothing parameters as the sequential implementation, the parallel implementation increases the smoothing effect by enlarging directly the value of σ G used in each scale.”; Wherein the input image is processed through the Gaussian smoothing filter sequentially.).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to implement the multicore processors, used to transform the input image in parallel, disclosed by Siriborvornratanakul, to perform the image augmentations disclosed by Hu in view of Wei. The suggestion/motivation for doing so would have been “Our experiments presented in Fig. 6 show that the parallel implementation offers similar multiscale detection outcomes compared with the sequential implementation. In this way, speed of detection can be improved significantly with few modifications.” (Siriborvornratanakul: Section III. D. Parallel implementation). Further, one skilled in the art could have combined the elements as described above by known methods with no change in their respective functions, and the combination would have yielded nothing more than predictable results. Therefore, it would have been obvious to combine Hu in view of Wei with Siriborvornratanakul to obtain the invention as specified in claim 3.
Regarding claim 11, Hu in view of Wei discloses: The system of claim 1.
Hu in view of Wei does not disclose expressly: wherein the one or more processors comprise one or more parallel processing units, and wherein the one or more parallel processing units apply the data transformation to the input data for multiple iterations in parallel.
Siriborvornratanakul discloses: wherein the one or more processors comprise one or more parallel processing units, and wherein the one or more parallel processing units apply the data transformation to the input data for multiple iterations in parallel (Siriborvornratanakul: Figure 6; Section III. D. Parallel implementation: “Considering recent growth in multicore processors and parallel programming languages, we changed the sequential implementation to the equivalent parallel implementation. The concept of our parallel implementation to the multiscale visual detection is illustrated in Fig. 6. The preprocessed image is distributed simultaneously to all scales. Instead of using the same smoothing parameters as the sequential implementation, the parallel implementation increases the smoothing effect by enlarging directly the value of σ G used in each scale.”; Wherein the input image is processed through the Gaussian smoothing filter sequentially.).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to implement the multicore processors, used to transform the input image in parallel, disclosed by Siriborvornratanakul, to perform the image augmentations disclosed by Hu in view of Wei. The suggestion/motivation for doing so would have been “Our experiments presented in Fig. 6 show that the parallel implementation offers similar multiscale detection outcomes compared with the sequential implementation. In this way, speed of detection can be improved significantly with few modifications.” (Siriborvornratanakul: Section III. D. Parallel implementation). Further, one skilled in the art could have combined the elements as described above by known methods with no change in their respective functions, and the combination would have yielded nothing more than predictable results. Therefore, it would have been obvious to combine Hu in view of Wei with Siriborvornratanakul to obtain the invention as specified in claim 11.
Regarding claim 20, Hu in view of Wei discloses: The processor of claim 17.
Hu in view of Wei does not disclose expressly: wherein the processor comprises parallel processing circuitry to accelerate performing the plurality of processing iterations with respect to the input data.
Siriborvornratanakul discloses: wherein a processor comprises parallel processing circuitry to accelerate performing a plurality of processing iterations with respect to the input data (Siriborvornratanakul: Figure 6; Section III. D. Parallel implementation: “Considering recent growth in multicore processors and parallel programming languages, we changed the sequential implementation to the equivalent parallel implementation. The concept of our parallel implementation to the multiscale visual detection is illustrated in Fig. 6. The preprocessed image is distributed simultaneously to all scales. Instead of using the same smoothing parameters as the sequential implementation, the parallel implementation increases the smoothing effect by enlarging directly the value of σ G used in each scale.”; Wherein the input image is processed through the Gaussian smoothing filter sequentially.).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to implement the multicore processors, used to transform the input image in parallel, disclosed by Siriborvornratanakul, to perform the image augmentations disclosed by Hu in view of Wei. The suggestion/motivation for doing so would have been “Our experiments presented in Fig. 6 show that the parallel implementation offers similar multiscale detection outcomes compared with the sequential implementation. In this way, speed of detection can be improved significantly with few modifications.” (Siriborvornratanakul: Section III. D. Parallel implementation). Further, one skilled in the art could have combined the elements as described above by known methods with no change in their respective functions, and the combination would have yielded nothing more than predictable results. Therefore, it would have been obvious to combine Hu in view of Wei with Siriborvornratanakul to obtain the invention as specified in claim 20.
Claim(s) 6-7 is rejected under 35 U.S.C. 103 as being unpatentable over Hu in view of Wei, and further in view of Goris et al. (US2021209764A1) hereinafter referenced as Goris.
Regarding claim 6, Hu in view of Wei discloses: The system of claim 4, wherein the one or more processors are further to: determine an incremental detection amount based at least on a comparison of the combined object mask with one or more existing object masks; and determine that the termination criteria has been satisfied based at least on the incremental detection amount (Hu: 0143: “the terminal device calculates an error between an original mask before the updating and an original mask after the updating, and detects whether the error is less than a preset error threshold. If the error is less than the preset error threshold, it indicates that the updated original mask satisfies the error convergence condition, and in this case, the terminal device may use the updated original mask as a target mask. If the error is not less than the preset error threshold, it indicates that the updated original mask does not satisfy the error convergence condition, and the original mask needs to be updated continuously and iteratively, until the updated original mask satisfies the error convergence condition.”).
Hu in view of Wei does not disclose expressly: wherein the one or more processors are further to: determine an incremental detection amount based at least on a comparison of the combined object mask with one or more existing object masks, the incremental detection amount indicating an amount of new object elements added to the combined object mask.
Goris discloses: wherein one or more processors configured to: determine an incremental detection amount based at least on a comparison of an object mask with one or more existing object masks, the incremental detection amount indicating an amount of new object elements added to the object mask (Goris: 0055: “In one embodiment, the total number of iterations depends on the extent to which the size of the segmentation mask is changing from one iteration to the next. For example, if the size of the segmentation mask is not changing by more than a threshold number of image elements, the iterative process 400 may be terminated. In other words, the iteration process 400 may continue until the difference between the number of image elements in the segmentation masks obtained in two successive iterations is smaller than a threshold value (e.g., a value comprised between 5 and 20).”; Wherein the difference between segmentation masks determines how much the current mask differs from the previous one, based on their respective image elements).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to substitute the error convergence condition disclosed by Hu in view of Wei with the image element based iteration termination process taught by Goris. The suggestion/motivation for doing so would have been “if the size of the segmentation mask is not changing by more than a threshold number of image elements, the iterative process 400 may be terminated. In other words, the iteration process 400 may continue until the difference between the number of image elements in the segmentation masks obtained in two successive iterations is smaller than a threshold value ( e.g., a value comprised between 5 and 20).” (Goris: 0055). Further, one skilled in the art could have substituted one known element for another, and the substitution would have yielded nothing more than predictable results. Therefore, it would have been obvious to combine Hu in view of Wei with Goris to obtain the invention as specified in claim 6.
Regarding claim 7, Hu in view of Wei discloses: The system of claim 4, wherein the one or more processors are further to: determine the difference between the combined object mask and one or more existing object masks (Hu: 0143: “the terminal device calculates an error between an original mask before the updating and an original mask after the updating, and detects whether the error is less than a preset error threshold. If the error is less than the preset error threshold, it indicates that the updated original mask satisfies the error convergence condition, and in this case, the terminal device may use the updated original mask as a target mask. If the error is not less than the preset error threshold, it indicates that the updated original mask does not satisfy the error convergence condition, and the original mask needs to be updated continuously and iteratively, until the updated original mask satisfies the error convergence condition.”).
Hu in view of Wei does not disclose expressly: wherein the one or more processors are further to: determine an incremental object mask as the difference between the combined object mask and one or more existing object masks.
Goris discloses: wherein the one or more processors are further to: determine an incremental object mask as the difference between an object mask and one or more existing object masks (Goris: 0055: “In one embodiment, the total number of iterations depends on the extent to which the size of the segmentation mask is changing from one iteration to the next. For example, if the size of the segmentation mask is not changing by more than a threshold number of image elements, the iterative process 400 may be terminated. In other words, the iteration process 400 may continue until the difference between the number of image elements in the segmentation masks obtained in two successive iterations is smaller than a threshold value (e.g., a value comprised between 5 and 20).”; Wherein the determining of the difference between the number of image elements in the segmentation masks constitutes the determining of a difference mask.).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to substitute the error convergence condition disclosed by Hu in view of Wei with the image element based iteration termination process taught by Goris. The suggestion/motivation for doing so would have been “if the size of the segmentation mask is not changing by more than a threshold number of image elements, the iterative process 400 may be terminated. In other words, the iteration process 400 may continue until the difference between the number of image elements in the segmentation masks obtained in two successive iterations is smaller than a threshold value ( e.g., a value comprised between 5 and 20).” (Goris: 0055). Further, one skilled in the art could have substituted one known element for another, and the substitution would have yielded nothing more than predictable results. Therefore, it would have been obvious to combine Hu in view of Wei with Goris to obtain the invention as specified in claim 7.
Claim(s) 11-12 are rejected under 35 U.S.C. 103 as being unpatentable over Wei, and further in view of Siriborvornratanakul.
Regarding claim 11, Wei discloses: The system of claim 1.
Wei does not disclose expressly: wherein the one or more processors comprise one or more parallel processing units, and wherein the one or more parallel processing units apply the data transformation to the input data for multiple iterations in parallel.
Siriborvornratanakul discloses: wherein the one or more processors comprise one or more parallel processing units, and wherein the one or more parallel processing units apply the data transformation to the input data for multiple iterations in parallel (Siriborvornratanakul: Figure 6; Section III. D. Parallel implementation: “Considering recent growth in multicore processors and parallel programming languages, we changed the sequential implementation to the equivalent parallel implementation. The concept of our parallel implementation to the multiscale visual detection is illustrated in Fig. 6. The preprocessed image is distributed simultaneously to all scales. Instead of using the same smoothing parameters as the sequential implementation, the parallel implementation increases the smoothing effect by enlarging directly the value of σ G used in each scale.”; Wherein the input image is processed through the Gaussian smoothing filter sequentially.).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to implement the multicore processors, used to transform the input image in parallel, disclosed by Siriborvornratanakul, to perform the image augmentations disclosed by Wei. The suggestion/motivation for doing so would have been “Our experiments presented in Fig. 6 show that the parallel implementation offers similar multiscale detection outcomes compared with the sequential implementation. In this way, speed of detection can be improved significantly with few modifications.” (Siriborvornratanakul: Section III. D. Parallel implementation). Further, one skilled in the art could have combined the elements as described above by known methods with no change in their respective functions, and the combination would have yielded nothing more than predictable results. Therefore, it would have been obvious to combine Wei with Siriborvornratanakul to obtain the invention as specified in claim 11.
Regarding claim 12, Wei in view of Siriborvornratanakul discloses: The system of claim 11, wherein the one or more parallel processing units are further to: perform the object detection on the new transformed data for multiple iterations in parallel (Wei: Figure 2; 3.1. Overview: “For the proposed system, during the in-line (testing) phase, the input goes through three stages. First, the input is augmented to produce several variations, so we can have augmented inputs for future stages. Second, a detector is applied to obtain AABBs and related labels in each augmented image. Practically, this would be implemented by applying multiple detectors in parallel, one for each of the augmented images…the system produces variations of each input, detects objects in each variation and fuses the top T results for each object, with the expectation of getting a more accurate AABB for each object in the input.”).
Conclusion
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 nonprovisional extension fee (37 CFR 1.17(a)) 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 mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ANTHONY J RODRIGUEZ whose telephone number is (703)756-5821. The examiner can normally be reached Monday-Friday 10am-7pm.
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, Sumati Lefkowitz can be reached at (571) 272-3638. 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.
/ANTHONY J RODRIGUEZ/Examiner, Art Unit 2672
/SUMATI LEFKOWITZ/Supervisory Patent Examiner, Art Unit 2672
Prosecution Insights