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 Amendment
The Amendment filed April 3 2026 has been entered and considered. Claims 1 and 11 have been amended. The amendment does not overcome the rejection under 35 U.S.C. 103 previously set forth. The rejection is maintained and is restated below with additional discussion of the amended limitations; accordingly, this action is made final.
Response to Arguments
Applicant's arguments filed 4/3/2026, Remarks Pgs. 7-8, have been fully considered but they are not persuasive.
Applicant argues that the prior art does not disclose the newly added amendments to the independent claims. Remarks of 4/3/26 at Pgs. 7-8. Examiner respectfully disagrees.
Applicant argues (Pg. 8):
For example, Applicant submits that Jung does not teach or suggest a displayed input mechanism or a range of probabilities included in the displayed input mechanism.
Examiner responds:
While Jung does not explicitly disclose a displayed input mechanism, the rejection under 35 U.S.C. 103 relies on the combination of Jung and the Roboflow software, not Jung alone. The Roboflow software explicitly teaches a displayed input mechanism (Nelson ‘21, Pg. 3, shows an example of the UI). Further, Jung explicitly discloses the ability for a user to direct a range of probabilities (Pg. 3, “…or decrease the probability of some augmenters to be applied by decreasing in (sometimes = lambda aug: iaa.Sometimes(0.5, aug) the value 0.5 to e.g. 0.3)”; Pg. 4, snippet provided,
PNG
media_image1.png
70
486
media_image1.png
Greyscale
). One of ordinary skill in the art would have understood that both the probability-based augmentation selection of Jung and the user interface augmentation system of the Roboflow software are directed to the same field of image augmentation for machine learning training sets. Since the Roboflow software already provides a user interface for configuring image augmentation operations, it would have been an obvious design choice to further implement the known probability-range parameter of Jung through the interface so users could adjust augmentation application likelihoods.
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, 7, 9-13, and 17-18 are rejected under 35 U.S.C. 103 as being unpatentable over the Roboflow software available as of 2021 as shown by Nelson (NPL, “Image Augmentations for Aerial Datasets”, dated 4/26/2021, pdf attached as well as video screenshots) hereinafter Nelson ‘21, and Gallagher (NPL, “Generate Image Augmentations with Roboflow”, published 2/24/2023, pdf attached) in view of Jung et al. (NPL, “Imgaug, Examples: Basics”, published 2020, pdf attached).
PNG
media_image2.png
510
715
media_image2.png
Greyscale
PNG
media_image3.png
582
980
media_image3.png
Greyscale
PNG
media_image4.png
489
657
media_image4.png
Greyscale
Regarding claim 1, the Roboflow software teaches a system comprising: a first computer including a processor coupled to a memory, the memory including instructions executable by the processor to (Pg. 5, “Training time is sped up because you don't spend compute resources waiting for your CPU to create augmented images during training time.”, a processor coupled to memory is necessary to utilize the system): obtain, via a displayed input mechanism, user input of parameters to direct probabilities included in the displayed input mechanism of modification types for a set of images to be modified from a base image stored in a database (Fig. on pg. 3, reprinted above, shows a plurality of possible modifications; Pg. 9, “For each image in the multiple, a random selection of your selected augmentations will be applied. For instance, if you have chosen crop, brightness, and resize augmentations with a 2x multiple, each image will have a random selection of those three augmentations applied.”, Also see screenshots above from the video included in Nelson ‘21 where a set of images are uploaded and then augmentation can be performed);
generate, responsive to the user input via the displayed input mechanism, representations of the set of images modified from the base image, the representations including the modifications according to the input parameters (Fig. on pg. 8, reprinted below, shows the base image being modified according to user input).
PNG
media_image5.png
340
452
media_image5.png
Greyscale
and transmit the representation of the base image to a second computer for display (Pg. 10, “With augmented images in a dataset version, you can kick off a training job on Roboflow to generate an infinitely-scalable API for use with our SDKs and web API, or you can export your dataset for use in custom training.”, an SDK and web API are disclosed, indicating the transmission of images and data between a server and client device, including the delivery of modified images from the server to the client).
Roboflow does not explicitly disclose obtaining user input to select respective probabilities, from a range of probabilities, of modification types for a set of images to be modified.
Jung teaches to obtain user input to select respective probabilities, from a range of probabilities, of modification types for a set of images to be modified from a base image stored in a database (Pg. 4, Code snippet, reprinted below,
PNG
media_image1.png
70
486
media_image1.png
Greyscale
, shows that the respective probability of a modification can be selected by the user. For example line 2 shows that the images have a 50% chance of being blurred; Fig. on pg. 3, reprinted below,
PNG
media_image6.png
262
520
media_image6.png
Greyscale
shows an example output dataset with randomly applied augmentations based on user selected probabilities).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Roboflow to incorporate the teachings of Jung to include obtaining user input to select respective probabilities, from a range of probabilities, of modification types for a set of images to be modified. The Roboflow framework teaches performing augmentation on images based on user selections, however, the frequency in which these modifications are applied to the image cannot be changed through user selection. Jung teaches to perform augmentation on images based on user selection, including allowing the user to provide a probability in which they would like the images to have an augmentation applied. One of ordinary skill in the art would find it obvious to parametrize the augmentation operations of Roboflow using the probabilistic augmentation techniques taught by Jung in order to allow user control over the frequency of each modification type. This would predictably increase the robustness of the system as the data will be more thoroughly and randomly augmented, reducing the chance that a neural network will overfit.
Regarding claim 2, the Roboflow software as modified teaches all of the elements of claim 1, as stated above, as well as wherein the respective probabilities comprise directions to: rotate the base image in a horizontal plane of the base image; rotate the base image in a vertical plane of the base image; modify contrast between portions of the base image; modify brightness of a portion of the base image; modify visual noise content of a portion of the base image; and/or add blur to a portion of the base image (Fig. on pg. 3, reprinted above, includes rotation, visual noise, brightness, and blur; Jung; Pg. 1, “It applies crops and affine transformations to images, flips some of the images horizontally, adds a bit of noise and blur and also changes the contrast as well as brightness.”, shows an example code snippet which can modify rotation, brightness, noise, and blur using user probabilities, Code snippet reprinted above, shows rotation; Pg. 2, Code snippet, reprinted below,
PNG
media_image7.png
28
332
media_image7.png
Greyscale
, shows contrast as being modifiable).
Regarding claim 4, the Roboflow software as modified teaches all of the elements of claim 1, as stated above, as well as wherein the user input parameters comprise a selection of a number of the representations of the base image to be transmitted to the second computer for display (Fig. on pg. 9, reprinted below, shows the capability for the user to select the
PNG
media_image8.png
263
510
media_image8.png
Greyscale
desired amount of representations).
Regarding claim 7, the Roboflow software as modified teaches all of the elements of claim 1, as stated above, as well as wherein the modified base image comprises an array of red-green-blue (RGB) values (Fig. on pg. 3, reprinted above, shows modified images comprising an array of RGB values).
PNG
media_image9.png
233
499
media_image9.png
Greyscale
Regarding claim 9, the Roboflow software as modified teaches all of the elements of claim 1, as stated above, as well as wherein transmission of the representations of the base image is substantially simultaneous with generation of the modified base image (Fig. on pg. 10, reprinted below, shows that generation takes a few moments before transmission is possible, indicating that generation is performed in near-real-time. Once generated, the images are immediately available for transmission via API, as taught by Gallagher. One of ordinary skill in the art would understand that transmission occurs substantially simultaneously with generation in such a real-time system).
Regarding claim 10, the Roboflow software as modified teaches all of the elements of claim 1, as stated above, as well as wherein the instructions further include instructions to: transmit the base image from the first computer to the second computer prior to obtaining the user input (Pg. 2, “The Roboflow platform provides a visual interface through which you can apply 23 different augmentations to a dataset.”; Pg. 10, “With augmented images in a dataset version, you can kick off a training job on Roboflow to generate an infinitely-scalable API for use with our SDKs and web API, or you can export your dataset for use in custom training.”, Augmentations are provided and can be applied, but they are not required. Accordingly, the API can be used for transmission of the base image prior to receiving user input for modification parameters, as the system allows for image handling both with and without augmentation).
Claim 11 corresponds to claim 1 and is rejected under the same analysis.
Claim 12 corresponds to claim 4 and is rejected under the same analysis.
Claim 13 corresponds to claim 2 and is rejected under the same analysis.
Claim 17 corresponds to claim 9 and is rejected under the same analysis.
Claim 18 corresponds to claim 10 and is rejected under the same analysis.
Claim(s) 3, 5, 8, 14, and 16 are rejected under 35 U.S.C. 103 as being unpatentable over the Roboflow software (as shown by Nelson ‘21 and Gallagher above) in view of Jung et al. further in view of Nelson (NPL, “What is Image Preprocessing and Augmentation?”, published 1/26/2020, pdf attached) hereinafter Nelson ‘20.
Regarding claim 3, the Roboflow software as modified in view of Jung teaches all of the elements of claim 1, as stated above. They do not explicitly disclose that the representations of the modified base image correspond to down-sampled modified representations of the base image. However, they do disclose performing a blur modification which is commonly used in combination with downsampling pipelines.
Nelson ‘20 teaches wherein the representations of the modified base image correspond to down-sampled modified representations of the base image (Pg. 4, “Moreover, input images may be various sizes, and some may be smaller than the desired input size. Best tips: preserving scale is not always required, filling in dead pixels with reflected image content is often best, and downsampling large images to smaller images is often safest.”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the Roboflow software and Jung to incorporate the teachings of Nelson ‘20 to include the representation of the modified base image corresponding to a down-sampled modified representation of the base image. Nelson ‘20 discloses downsampling within the same Roboflow framework used by Nelson ‘21 and Gallagher. Although described as a preprocessing step rather than an augmentation, one of ordinary skill in the art would recognize that including downsampling in the user-directed image modification process would provide additional control parameters for affecting the visual characteristic of the image and contribute to a more robust defense against overfitting.
Regarding claim 5, the Roboflow software as modified in view of Jung teaches all of the elements of claim 1, as stated above. The additional limitations are taught in combination with Nelson ‘20, which discloses wherein the representations of the modified base image correspond to down-sampled modified representations of the base image (See analysis above), wherein the modified representations are down-sampled by a user-selectable amount (The Roboflow software discloses user-selectable amounts for numerous modifications. One of ordinary skill in the art would recognize that extending this functionality to include a user-selectable downsampling amount would be an obvious design choice providing users with flexibility and maintaining consistency within the existing modification framework).
Regarding claim 8, the Roboflow software as modified in view of Jung teaches all of the elements of claim 1, as stated above. The additional limitations are taught in combination with Nelson ‘20, which discloses wherein the representations of the modified base image correspond to down-sampled modified representations of the base image, and wherein the instructions executable by the processor further include instructions to: obtain a down sampling value to be applied to the modified base image (See analysis for claim 5 above. For a user-selectable downsampling amount to be applied, the processor would necessarily obtain a downsampling value, whether from user input, configuration settings, or another input mechanism).
Claim 14 corresponds to claim 3 and is rejected under the same analysis.
Claim 16 corresponds to claim 5 and is rejected under the same analysis.
Claim 6 is rejected under 35 U.S.C. 103 as being unpatentable over the Roboflow software as shown by Nelson ‘21 and Gallagher in view of Jung et al. further in view of Robbins (NPL, “A beginner's guide to JSON, the data format for the internet”, published 6/2/2022, pdf attached).
Regarding claim 6, the Roboflow software as modified in view of Jung teaches all of the elements of claim 1, as stated above. They do not explicitly disclose using JSON for the transmission process; however, they do perform data transmission through APIs.
Robbins teaches wherein the representations of the modified base image are to be transmitted in a JavaScript Object Notation (JSON) format (Pg. 3, “One of the most common uses for JSON is when using an API, both in requests and responses. It is much more compact than other standards and allows for easy consumption in web browsers as JavaScript can easily parse JSON strings, only requiring JSON.parse() to start using it.”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the Roboflow software to incorporate the teachings of Robbins to include wherein the representation of the modified base image is to be transmitted in a JavaScript Object Notation (JSON) format. The Roboflow software discloses using a web based API for transmission between devices. Robbins discloses that “when APIs send data, chances are they send it as JSON objects” (Pg. 1). One of ordinary skill in the art would know of the JSON format and its advantages, such as human readability and compatibility with JavaScript based websites, leaving it an obvious choice in view of Robbins (Pg. 3).
Claim 15 is rejected under 35 U.S.C. 103 as being unpatentable over the Roboflow software as shown by Nelson ‘21 and Gallagher in view of Jung et al. further in view of Dodge (NPL, “Understanding How Image Quality Affects Deep Neural Networks”, published 4/21/2016, pdf attached).
Regarding claim 15, the Roboflow software as modified in view of Jung teaches all of the elements of claim 11, as stated above. They do not explicitly disclose the representations of the modified base image correspond to compressed modified representations of the base image.
Dodge teaches the representations of the modified base image correspond to compressed modified representations of the base image (Pg. 3, “For JPEG compression, in our experiments we vary the quality parameter from 2 to 20 in steps of 2. A quality parameter value of 100 represents the original uncompressed image”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the Roboflow software to incorporate the teachings of Dodge to include the representation of the modified base image corresponds to a compressed modified representation of the base image. Compression is a well-known processing step used to reduce file size. One of ordinary skill in the art would recognize that “compression is interesting to study because if equivalent performance can be achieved at higher compression ratios, the storage or memory requirement for certain applications can be reduced” (Dodge, Pg. 3). Dodge also discloses that “in initial experiments, we found that the accuracy of the networks does not significantly decrease between quality levels 100 to 20”, showcasing the advantages of compressing images for machine learning processing.
Conclusion
THIS ACTION IS MADE FINAL. 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 DAVID A WAMBST whose telephone number is (703)756-1750. The examiner can normally be reached M-F 9-6:30 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, Gregory Morse can be reached at (571)272-3838. 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.
/DAVID ALEXANDER WAMBST/Examiner, Art Unit 2663
/GREGORY A MORSE/Supervisory Patent Examiner, Art Unit 2698
Prosecution Insights