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 with respect to claims 1-20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
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.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Appalaraju et al. (Appalaraju) (US 10,467,526 B1).
Regarding claim 1, Appalaraju discloses a system (e.g., system environment in which a combination of curriculum learning based-training data set selection and a multi-scale convolutional neural network model may be employed for image similarity analysis, paragraph 1, figure 1), comprising:
a storage medium (e.g., image source 110, image collectors 112, figure 1); and
a processor comprising circuitry (e.g., In some embodiments, at least some combinations of the elements of system 100 that are shown as separate entities in FIG. 1 (such as training coordinators, run-time coordinators, image collectors, image pair selection subsystems, model training resources, pre-processors, model execution resources and the like) may be implemented using a single computing device, paragraph 32) configured to:
generate, from a neural network for processing a plurality of inputs corresponding to a set of color channels in instances of input image data, a sub-network for processing a reduced set of the plurality of inputs that includes fewer color channels than the set of color channels to be processed by the neural network (e.g., in at least some embodiments, a multi-scale neural network model in which different numbers of convolution layers are configured for respective subnetworks may be employed. FIG. 3 illustrates an examples of a multi-scale convolutional neural network which may be used at each of the mirror networks of an image similarity analysis model, according to at least some embodiments. As shown, multi-scale CNN 320 may comprise at least three convolution subnetworks in the depicted embodiment: subnetworks 351, 352 and 353, with 16, 3 and 2 convolution layers respectively, paragraph 36); and 36, …526
store the sub-network in the storage medium (e.g., In one embodiment, a collection of images (or image pairs) as well as intermediate and/or final results of analysis of the images (such as the kinds of vectors produced using neural network models discussed above) may be stored at a storage service 823, paragraph 88).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to have modified Appalaraju in one embodiment to include store the sub-network in the storage medium in another embodiment to use for a particular application.
Regarding claim 2, Appalaraju discloses wherein the processor is configured to generate, from the neural network, an additional sub-network for processing an additional set of color channels that includes: fewer color channels than the set of color channels; and different color channels than the reduced set of color channels (e.g., FIG. 3 illustrates an examples of a multi-scale convolutional neural network which may be used at each of the mirror networks of an image similarity analysis model, according to at least some embodiments. As shown, multi-scale CNN 320 may comprise at least three convolution subnetworks in the depicted embodiment: subnetworks 351, 352 and 353, with 16, 3 and 2 convolution layers respectively, paragraph 36).
Regarding claim 3, Appalaraju discloses wherein the processor is configured to combine the sub-network and the additional sub-network to create an ensemble network (e.g., subnetworks 352, 353, figure 3).
Regarding claim 4, Appalaraju discloses wherein the processor is configured to select the sub-network and/or the additional sub-network for processing instances of input image data based at least in part on one or more color channels to be processed (e.g., subnetworks 351, 352, 353, figure 3).
Regarding claim 5, Appalaraju discloses wherein the processor is configured to provide representations of the sub-network and the additional sub-network to an additional processor, the representations of the sub-network and the additional sub-network configured to be combined by the additional processor into an ensemble network to be used for processing instances of input image data (e.g., In at least some embodiments, the subnetworks may generate output vectors or embeddings of different lengths. For example, as shown, subnetwork 351 may emit a 4096-element vector embedding, subnetwork 352 may produce a 1024-element vector, while subnetwork 353 may generate a 512-element vector, with each of the vectors representing features of the same input image as learned by the respective subnetworks. The CNN 320 may be referred to as a multi-scale CNN because convolution at different scales (using respective numbers of convolution layers) may be performed on the same input image, resulting in different output feature vectors which may have to be aggregated or combined to generate the final output of the model as a whole, paragraph 40).
Regarding claim 6, Appalaraju discloses wherein the processor is configured to provide a representation of the sub-network to an additional processor, the representation of the sub-network configured to be used by the additional processor for processing instances of input image data (e.g., the configuration and sizes for various components and artifacts of a multi-scale CNN used in various embodiments may differ from those shown by way of example in FIG. 3. For example, the sizes of the input image representations, the number of subnetworks, the number of convolution layers at individual ones of the subnetworks, the sampling ratios (if sampling is used at all), the lengths of the vectors, and/or the number and types of aggregation/normalization layers may differ in some embodiments from those shown in FIG. 3, paragraph 41).
Regarding claim 7, Appalaraju discloses wherein generating the sub-network includes creating a copy of the neural network from which internal elements for processing at least one color channel are removed (e.g., For example, sub-sampler 312A may use a 4:1 sampling ratio to convert an input image 310 comprising 224×224 pixels into a 56×56 pixel (which are color channels are removed) sample used as input to the subnetwork 352, while sub-sampler 312B may use a 8:1 sampling ratio to convert the 224×224 pixels into a 28×28 pixel (which are color channels are removed) sample used as input to the subnetwork 353, paragraph 39).
Regarding claim 8, claim 8 is a method claim with limitations similar of limitations of claim 1. Therefore, claim 8 is rejected as set forth above as claim 1.
Regarding claim 9, claim 9 is a method claim with limitations similar of limitations of claim 2. Therefore, claim 9 is rejected as set forth above as claim 2.
Regarding claim 10, claim 10 is a method claim with limitations similar of limitations of claim 3. Therefore, claim 10 is rejected as set forth above as claim 3.
Regarding claim 11, claim 11 is a method claim with limitations similar of limitations of claim 4. Therefore, claim 11 is rejected as set forth above as claim 4.
Regarding claim 12, claim 12 is a method claim with limitations similar of limitations of claim 5. Therefore, claim 12 is rejected as set forth above as claim 5.
Regarding claim 13, claim 13 is a method claim with limitations similar of limitations of claim 6. Therefore, claim 13 is rejected as set forth above as claim 6.
Regarding claim 14, claim 14 is a method claim with limitations similar of limitations of claim 7. Therefore, claim 14 is rejected as set forth above as claim 7.
Regarding claim 15, claim 15 is a processor claim with limitations similar of limitations of claim 1. Therefore, claim 15 is rejected as set forth above as claim 1.
Regarding claim 16, claim 16 is a processor claim with limitations similar of limitations of claim 2. Therefore, claim 16 is rejected as set forth above as claim 2.
Regarding claim 17, claim 17 is a processor claim with limitations similar of limitations of claim 3. Therefore, claim 17 is rejected as set forth above as claim 3.
Regarding claim 18, claim 18 is a processor claim with limitations similar of limitations of claim 4. Therefore, claim 18 is rejected as set forth above as claim 4.
Regarding claim 19, claim 19 is a processor claim with limitations similar of limitations of claim 5. Therefore, claim 19 is rejected as set forth above as claim 5.
Regarding claim 20, claim 20 is a processor claim with limitations similar of limitations of claim 6. Therefore, claim 20 is rejected as set forth above as claim 6.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to QUANG N VO whose telephone number is (571)270-1121. The examiner can normally be reached Monday-Friday, 7AM-4PM, EST.
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, Abderrahim Merouan can be reached at 571-270-5254. 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.
/QUANG N VO/Primary Examiner, Art Unit 2683
