SYSTEM AND METHOD FOR AUTONOMOUS BOTANICAL IDENTIFICATION AND ADULTERATION DETECTION

§103 §112

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 . Claim Objections Claim 1 is objected to because of the following informalities: In line 2, “the system comprising” should be -the botanical identification and adulteration detection system comprising- In line 8, “the raw HPTLC images” should be -the plurality of raw HPTLC images- In line 17, “synthetic data” should be -the HPTLC synthetic data- In line 20, “the HPTLC GAN discriminator machine learning model” should be -an HPTLC GAN discriminator machine learning model- In line 26, “memory storing a synthetic dataset containing synthetic data generated by the GAN component” should be -the memory storing the synthetic dataset generated by the GAN generator component- In line 37, “the botanical labels” should be -the botanical material labels-. Appropriate correction is required. Claim 3 is objected to because of the following informalities: “the user interface” should be -a user interface-. Appropriate correction is required. Claim 10 is objected to because of the following informalities: “include” should be -includes-. Appropriate correction is required. Claim 11 is objected to because of the following informalities: In line 2, “the method including” should be -the method for identifying botanical content and detecting adulteration including- In line 7, “the raw HPTLC images” should be -the plurality of raw HPTLC images- In line 16, “synthetic data” should be -the HPTLC synthetic data- In line 17, “the HPTLC GAN discriminator machine learning model” should be -a HPTLC GAN discriminator machine learning model- In line 25, “memory storing a synthetic dataset containing synthetic data generated by the GAN component” should be -the memory storing the synthetic dataset generated by the GAN generator component- In line 35, “the botanical labels” should be -the botanical material labels- Appropriate correction is required. Claim Rejections - 35 USC § 112 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. Claims 1-20 are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor regards as the invention Claims 6 and 7 recites “the test dataset”. Claim 1 has a GAN test data set and a CNN test data set. It is unclear which data set the test dataset refers to. Therefore claims 6-7 are indefinite. 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. The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows: 1. Determining the scope and contents of the prior art. 2. Ascertaining the differences between the prior art and the claims at issue. 3. Resolving the level of ordinary skill in the pertinent art. 4. Considering objective evidence present in the application indicating obviousness or nonobviousness. Claims 1-2, 6, 8-11, 17, and 19-20 are rejected under 35 USC 103 as being unpatentable over Jahanbakhshi et al. (Detection of Fraud in Ginger Powder Using an Automatic Sorting system Based on Image Processing Technique and Deep Learning, September 2021, Elsevier; hereinafter Jahanbakhshi) in view of Debnath et al. (US 20250151639, hereinafter Debnath), further in view of Pieszczek et al. (Multi-Wavelength Imaging of HPTLC Plates Using a Constructed Illumination Chamber with a Smartphone Camera as the Detector, January 2021, Elsevier; hereinafter Pieszczek), further in view of Wu et al. (US 20240419859, hereinafter Wu). Claims 1 and 11 Jahanbakhshi teaches a botanical identification and adulteration detection system and corresponding method (Abstract — “This study classified ginger powder images to detect fraud by improving convolutional neural networks (CNN) through a gated pooling function.”) for use with output images (Abstract — “For this purpose, 3360 image samples of ginger powder were prepared in 7 categories…”), the system comprising: memory storing a plurality of images (Section 2.5 — “In the proposed method, images of ginger powder were classified using CNN. In the training stage, images of ginger powder were fed to the network.”; a CNN receiving images discloses the existence of a computer with memory from which to take the images) and associated labels (Abstract — “For this purpose, 3360 image samples of ginger powder were prepared in 7 categories (pure ginger powder, chickpea powder, 10%, 20%, 30%, 40%, and 50% fraud in ginger powder)”; the categories are labels), the associated labels including one or more botanical material labels and one or more adulterant labels (Abstract — “For this purpose, 3360 image samples of ginger powder were prepared in 7 categories (pure ginger powder, chickpea powder, 10%, 20%, 30%, 40%, and 50% fraud in ginger powder)”); one or more computer subsystems (Section 2.5 — “In the proposed method, images of ginger powder were classified using CNN. In the training stage, images of ginger powder were fed to the network.”; training a CNN discloses the existence of a computer with a processor to perform the training); and one or more components executed by the one or more computer subsystems, wherein the one or more components include: an image processing component configured to convert the raw images into a dataset, including a test dataset and a validation dataset (Section 2.4, Table 1 — The images are converted into train, test, and validation datasets.”), each of the test dataset and the validation dataset including a plurality of features (the images contain features, see Section 2.6 — “…the desired properties were first extracted from 3360 color images of ginger powder (without data augmentation) by the histograms of oriented gradients (HOG) and local binary pattern (LBP).”); a convolutional neural network (CNN) training component (Section 2.5 — “In the training stage, images of ginger powder were fed to the network. Then, through an error function (softmax), the output of CNN network was calculated. The network output was compared with the correct response, and the error rate was calculated to adjust the network parameters.”; having trained a CNN discloses the existence of a component to perform the training) configured to receive a CNN test dataset and a CNN validation dataset (Section 3, Table 4 — The table has performance metrics for training, testing, and validation, thereby disclosing the CNN training component receiving the test and validation datasets), each of the CNN test dataset and the CNN validation dataset including at least a portion of the image dataset (Section 2.4, Table 1 — The table shows the CNN test dataset and the CNN validation dataset portioned from the image dataset), the CNN training component configured to train a CNN machine learning model as a function of the CNN test dataset (Section 2.4, Table 1 and Section 3, Table 4 — That the data is portioned into train, test, and validation subsets means the training is influenced by the test dataset and its performance metrics), the CNN training component includes a CNN validation component configured to validate the CNN machine learning model as a function of the validation dataset (Section 2.4, Table 1 and Section 3, Table 4 — That the data is portioned into train, test, and validation subsets means the training is influenced by the validation dataset and its performance metrics and that there exists a validation component for computing said metrics); and a CNN processing component (Section 3, Table 4 — That these performance metrics exist discloses the existence of a CNN processing component used to compute them) configured to receive a new image dataset representing an image, and to process the new image dataset to identify botanical material associated with one or more of the botanical labels and to detect adulterants associated with one or more of the adulterant labels using the validated CNN machine learning model (Abstract — “Therefore, it can be said that the CNN method and image processing technique effectively increase marketability, prevent ginger powder fraud, and promote traditional methods of ginger powder fraud detection.”; this discloses the ability of the CNN to be used on new images in new datasets from real-world applications, identifying botanical material and detecting adulterants as it does in the training process). Jahanbakhshi fails to teach converting the images into a tensor dataset, High-Performance Thin-Layer Chromatography (HPTLC) images, and a generative adversarial network (GAN) component configured to generate a synthetic dataset, the GAN component including: a noise generator component configured to generate noise; a GAN generator component configured to produce HPTLC synthetic data as a function of an HPTLC GAN generator machine learning model and the noise generated by the noise generator; a GAN discriminator component configured to evaluate synthetic data produced by the GAN generator as a function of the HPTLC GAN discriminator machine learning model and the GAN test dataset; a GAN training component configured to train the HPTLC GAN generator machine learning model and the HPTLC GAN discriminator machine learning model based on the evaluation; and a GAN validation component configured to validate the HPTLC GAN generator machine learning model and the HPTLC GAN discriminator machine learning model against the GAN validation dataset; memory storing a synthetic dataset containing synthetic data generated by the GAN component using the validated HPTLC machine learning GAN model. Debnath teaches converting the images into a tensor data ([0034] — “For example, the processing circuit 20 may use a so-called numpify routine to convert the first image data into NumPy arrays, usable by the Python library for working with arrays. Once converted, the processing circuit 20 performs a frame-to-transformation operation (using Python library operations) on the converted first image data to generate first image tensor data.”). It would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, to include the teachings of Debnath of converting the images into a tensor dataset into the invention of Jahanbakhshi for the purpose of enabling accurate image-based control to better prepare the data for use in training neural networks. Jahanbakhshi and Debnath fails to teach High-Performance Thin-Layer Chromatography (HPTLC) images and a generative adversarial network (GAN) component configured to generate a synthetic dataset, the GAN component including: a noise generator component configured to generate noise; a GAN generator component configured to produce HPTLC synthetic data as a function of an HPTLC GAN generator machine learning model and the noise generated by the noise generator; a GAN discriminator component configured to evaluate synthetic data produced by the GAN generator as a function of the HPTLC GAN discriminator machine learning model and the GAN test dataset; a GAN training component configured to train the HPTLC GAN generator machine learning model and the HPTLC GAN discriminator machine learning model based on the evaluation; and a GAN validation component configured to validate the HPTLC GAN generator machine learning model and the HPTLC GAN discriminator machine learning model against the GAN validation dataset; memory storing a synthetic dataset containing synthetic data generated by the GAN component using the validated HPTLC machine learning GAN model. Pieszczek teaches raw High-Performance Thin-Layer Chromatography (HPTLC) images (Section 3.1 — “Mixtures of different commercially available inks were chromatographed using HPTLC silica gel 60 plates (20 × 20 cm)…”; Section 3.5, Fig. 3 — This discloses an apparatus for acquiring HPTLC images from the raw measurements). It would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, to use raw HPTCL images as taught by Pieszczek in the system and datasets taught by Jahanbakhshi to more-accurately and more-precisely understand the composition of adulterants in a botanical sample. Together, the combination of Jahanbakhshi, Debnath, and Pieszczek still fails to teach a generative adversarial network (GAN) component configured to generate a synthetic dataset, the GAN component including: a noise generator component configured to generate noise; a GAN generator component configured to produce HPTLC synthetic data as a function of an HPTLC GAN generator machine learning model and the noise generated by the noise generator; a GAN discriminator component configured to evaluate synthetic data produced by the GAN generator as a function of the HPTLC GAN discriminator machine learning model and the GAN test dataset; a GAN training component configured to train the HPTLC GAN generator machine learning model and the HPTLC GAN discriminator machine learning model based on the evaluation; and a GAN validation component configured to validate the HPTLC GAN generator machine learning model and the HPTLC GAN discriminator machine learning model against the GAN validation dataset; memory storing a synthetic dataset containing synthetic data generated by the GAN component using the validated HPTLC machine learning GAN model. Wu teaches a generative adversarial network (GAN) component ([0047] — “In embodiments of the disclosure, the simulation system 110A is implemented as a GAN architecture…”) configured to generate a synthetic dataset ([0047] — “The generator's goal is to produce realistic representations of the cleaning efficacy given a set of input parameters… while the discriminator's goal is to differentiate between the true experimental data and the generated data.”), the GAN component including: a noise generator component configured to generate noise ([0050] — “…the training of the GAN architecture of the simulation system 110A can include providing random noise 230 to the generator 210.”; this discloses the existence of a component for generating noise); a GAN generator component ([0047] — “…the simulation system 110A is implemented as a GAN architecture having a generator/PINN 210 and a discriminator network 220.”) configured to produce synthetic data as a function of a GAN generator machine learning model ([0047] — “The generator's goal is to produce realistic representations of the cleaning efficacy…”; the generator is a GAN generator machine learning model) and the noise generated by the noise generator ([0050] — “…the training of the GAN architecture of the simulation system 110A can include providing random noise 230 to the generator 210.”); a GAN discriminator component ([0047] — “…the simulation system 110A is implemented as a GAN architecture having a generator/PINN 210 and a discriminator network 220.”) configured to evaluate synthetic data produced by the GAN generator as a function of the GAN discriminator machine learning model ([0047] — “…the discriminator's goal is to differentiate between the true experimental data and the generated data.”) and the GAN test dataset ([0047] — “…the discriminator's goal is to differentiate between the true experimental data and the generated data.”; the true experimental data is a dataset for testing the GAN during training); a GAN training component ([0050] — “The training of the GAN architecture of the simulation system 110A can be performed in accordance with embodiments of the disclosure…”) configured to train the GAN generator machine learning model and the GAN discriminator machine learning model ([0050] — “The generator network 210 and the discriminator network 220 are trained substantially concurrently…”) based on the evaluation ([0050] — “The training process continues until the generator 210 can produce realistic predictions of the output(s)…”); and a GAN validation component ([0050] — “The trained GAN architecture of the simulation system 100A can be validated…”; this discloses the existence of a component for validating the GAN) configured to validate the GAN generator machine learning model and the GAN discriminator machine learning model against the GAN validation dataset ([0050] — “The trained GAN architecture of the simulation system 100A can be validated using a separate dataset not used during training to assess the performance of the models of the simulation system 110A by comparing its predictions to the actual output(s) [observed] in the validation dataset.”); memory storing a synthetic dataset containing synthetic data generated by the GAN component using the validated machine learning GAN model ([0047] — “The generator's goal is to produce realistic representations of the cleaning efficacy given a set of input parameters…”; a person having ordinary skill in the art would have understood the synthetic data is stored in the memory of the computer system in Fig. 9). It would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, to use a GAN as taught by Wu to generate synthetic data for use in the CNN test and validation datasets in training the CNN taught by the combination of Jahanbakhshi, Debnath, and Pieszczek to improve CNN model performance in accurately detecting adulterants in botanical samples. Claim 2 The examiner respectfully points out that the combination further teaches including a user interface configured to output the results of the CNN processing component (Jahanbakhshi, Section 3, Fig. 11 — The outputting of this results matrix discloses the existence of a user interface for such outputting). Claim 6 The examiner respectfully points out that the combination further teaches wherein the test dataset includes multi-channel data representing red, green and blue pixel data (Pieszczek, Section 1 — “They are collected using UV-A, UV-C and LED illumination sources, and the resulting images are analyzed using three different detection channels of a smartphone camera sensor (three color components: red (R), green (G) and blue (B))”). Claim 8 The examiner respectfully points out that the combination further teaches further including an HPTLC system for producing HPTLC output from a target sample (Pieszczek, Section 3.1 — “Mixtures of different commercially available inks were chromatographed using HPTLC silica gel 60 plates (20 × 20 cm) as the stationary phase aluminum (Merck KgaA, Darmstadt, Germany).”; Pieszczek, Section 3.4 — “The densitograms were collected using a CD60 HPTLC instrument (Desaga GmbH, Wiesloch, Germany).”). Claim 9 The examiner respectfully points out that the combination further teaches wherein the HPTLC system includes an HPTLC plate (Pieszczek, Section 3.1 — “Mixtures of different commercially available inks were chromatographed using HPTLC silica gel 60 plates (20 × 20 cm) as the stationary phase aluminum (Merck KgaA, Darmstadt, Germany).”). Claim 10 The examiner respectfully points out that the combination further teaches wherein the HPTLC system include an image capture device configured to capture images of the HPTLC plate (Pieszczek, Section 3.5, Fig. 3 — This discloses an apparatus for acquiring raw HPTLC images from the plate). Claim 17 The examiner respectfully points out that that the combination further teaches wherein the test dataset include multi-channel data representing red, green and blue pixel data (Pieszczek, Section 1 — “They are collected using UV-A, UV-C and LED illumination sources, and the resulting images are analyzed using three different detection channels of a smartphone camera sensor (three color components: red (R), green (G) and blue (B))”). Claim 19 The examiner respectfully points out that the combination further teaches including the step of preforming HPTLC on a target sample to obtain an HPTLC plate (Pieszczek, Section 3.1 — “Mixtures of different commercially available inks were chromatographed using HPTLC silica gel 60 plates (20 × 20 cm) as the stationary phase aluminum (Merck KgaA, Darmstadt, Germany).”). Claim 20 The examiner respectfully points out that the combination further teaches including the step of capturing an image of the HPTLC plate (Pieszczek, Section 3.5, Fig. 3 — This discloses an apparatus for acquiring raw HPTLC images from the plate); and at least one of the following: providing the captured image as a raw HPTLC image (Pieszczek, Section 3.5, Fig. 3 — This discloses an apparatus for acquiring raw HPTLC images from the plate) to become part of the tensor dataset (the combination already discloses a tensor dataset made of raw HPTLC images); and providing the captured image as a new image dataset to the CNN processing component (the combination already discloses using the CNN on a new image dataset; the raw HPTLC image would be used to make the new dataset). Claims 3-4 and 12-13 are rejected under 35 USC 103 as being unpatentable over the combination of Jahanbakhshi, Pieszczek, Debnath, and Wu in view Grant et al. (US 20210092891 A1, hereinafter Grant). Claim 3 The examiner respectfully points out that the combination teaches wherein the use interface is configured to output results (Jahanbakhshi, Section 3, Fig. 11 — The outputting of this results matrix discloses the existence of a user interface for such outputting). However, the combination fails to teach at least one of a genus and a species. Grant teaches a machine learning model for identifying genus and species ([0066] — “In some such implementations, a different machine learning model may be trained for each distinct type of plant (e.g., genus, species)”). It would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, to use the model as taught by Grant to identify the genus and species of a compound and output it with the user interface taught by the combination to better specify to the user the composition of the botanical sample. Claim 4 The examiner respectfully points out that the combination with Grant further teaches wherein the user interface is configured to output at least one probability determination of that at least one of a genus and a species (Grant, [0066] — “In other implementations, a single machine learning model or an ensemble of machine learning models may be trained to generate multiple outputs, with each output representing a probability that vision data 411 depicts a different type of plant.”). Claim 12 The examiner respectfully points out that the combination with Grant further teaches further including the step of outputting at least one of a genus and a species of a botanical material identified in the processing step (Jahanbakhshi, Section 3, Fig. 11 — This figure shows the outputting of a results matrix; [0066] — “In some such implementations, a different machine learning model may be trained for each distinct type of plant (e.g., genus, species)”). Claim 13 The examiner respectfully points out that the combination with Grant further teaches wherein the step of outputting further includes outputting a probability associated the identified botanical material (Grant, [0066] — “In other implementations, a single machine learning model or an ensemble of machine learning models may be trained to generate multiple outputs, with each output representing a probability that vision data 411 depicts a different type of plant.”). Claim 5 is rejected under 35 USC 103 as being unpatentable over the combination of Jahanbakhshi, Pieszczek, Debnath, and Wu in view of Puttagunta, Sravan (US 20220188973 A1, hereinafter Puttagunta). Claim 5 The combination fails to teach wherein the one or more computer subsystems include an interpolation component to generate supplemental synthetic data using an interpolation algorithm. However, Puttagunta teaches an interpolation component to generate supplemental synthetic data using an interpolation algorithm ([0005] — “Aspects may include a method of augmenting camera devices with neural networks to control framerate, time synchronization, image stitching and resolution, the method comprising creating synthetic camera frames by using neural networks to interpolate between actual frame captures”; that the interpolation is being performed discloses the existence of an interpolation component). It would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, to include an interpolation component to generate supplemental synthetic data using an interpolation algorithm as taught by Puttagunta to reduce bias from the GAN in the synthetic dataset by creating synthetic data without using the GAN. Claims 7 and 18 are rejected under 35 USC 103 as being unpatentable over the combination of Jahanbakhshi, Pieszczek, Debnath, and Wu in view of Kolatt, Tsafrir (US 20250272840 A1, hereinafter Kolatt). Claim 7 The examiner respectfully points out that the combination teaches wherein the test dataset includes multi-channel data for red, green, and blue pixel data (Pieszczek, Section 1 — “They are collected using UV-A, UV-C and LED illumination sources, and the resulting images are analyzed using three different detection channels of a smartphone camera sensor (three color components: red (R), green (G) and blue (B))”). The combination fails to teach wherein the test dataset includes multi-channel data with values 0 to 255 for each of the red, green and blue pixel data. Kolatt teaches multi-channel data with values 0 to 255 for each of the red, green and blue pixel data ([0102] — “As appreciated, in a RGB color model, each color is represented as a combination of three primary colors: red, green, and blue values, usually ranging, for example, from 0 to 255 such that for example, red is represented as (255, 0, 0), green is (0, 255, 0), and blue is (0, 0, 255).”). It would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, to use the usual RGB color pixels values ranging from 0 to 255 taught by Kolatt in order to be consistent with the most common value range for RGB pixels. Claim 18 The examiner respectfully points out that the combination with Kolatt further teaches wherein the test dataset includes multi-channel data (Pieszczek, Section 1 — “They are collected using UV-A, UV-C and LED illumination sources, and the resulting images are analyzed using three different detection channels of a smartphone camera sensor (three color components: red (R), green (G) and blue (B))”) with values 0 to 255 for each of the red, green and blue pixel data (Kolatt, [0102] — “As appreciated, in a RGB color model, each color is represented as a combination of three primary colors: red, green, and blue values, usually ranging, for example, from 0 to 255 such that for example, red is represented as (255, 0, 0), green is (0, 255, 0), and blue is (0, 0, 255).”). Claims 14-15 are rejected under 35 USC 103 as being unpatentable over the combination in view of Jiang et al. (Metabolic Profiling and Phylogenetic Analysis of Medicinal Zingiber Species: Tools for Authentication of Ginger (Zingiber Officinale Rosc.), August 2006, Elsevier; hereinafter Jiang). Claim 14 The examiner respectfully points out that the combination further teaches wherein the step of storing in memory a plurality of raw HPTLC images and associated labels includes storing in memory a plurality of raw HPTLC images and associated labels of a single target botanical (Jahanbakhshi, Abstract — “This study classified ginger powder images to detect fraud by improving convolutional neural networks (CNN) through a gated pooling function.”; the only botanical target is ginger). The combination fails to teach wherein the step of storing in memory a plurality of raw HPTLC images and associated labels includes storing in memory a plurality of HPTLC images and associated labels of a plurality of common adulterants associated with the single target species. Jiang teaches a plurality of common adulterants associated with the single target species, ginger (Section 2.1 — “We, therefore, used solely the rpsl6 and trnL-F regions for our analysis of Z. officinale, Z. mioga, Z. montanum, Z. spectabile, Z. zerumbet and Alpinia galanga. All species were chosen because of their traditional use as medicinal plants and/or because they are used as adulterants of ginger…”). It would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, to take the plurality of common adulterants taught by Jiang and take HPTLC images and store them in memory with their respective associated labels as taught by the combination in order to improve adulterant detection by checking for known common adulterants in the botanical sample. Claim 15 The examiner respectfully points out that the combination with Jiang further teaches wherein the step of storing in memory a plurality of raw HPTLC images and associated labels includes storing in memory a plurality of raw HPTLC images and associated labels of Zingiber offinicale (Jahanbakhshi, Abstract — “This study classified ginger powder images to detect fraud by improving convolutional neural networks (CNN) through a gated pooling function.”; Zingiber Offinicale is ginger) and a plurality of HPTLC images and associated labels ofat least two of the following adulterants: Alpinia officinarum, Boesenbergia rotunda, Kaempferia galanga, Kaempferia parviflora, Zingiber montanum and Zingiber zerumbet (Section 2.1 — “We, therefore, used solely the rpsl6 and trnL-F regions for our analysis of Z. officinale, Z. mioga, Z. montanum, Z. spectabile, Z. zerumbet and Alpinia galanga. All species were chosen because of their traditional use as medicinal plants and/or because they are used as adulterants of ginger…”). Claim 16 is rejected under 35 USC 103 as being unpatentable over the combination of Jahanbakhshi, Pieszczek, Debnath, and Wu in view of Puttagunta and in further view of Lott, Steven (US 20210191908 A1, hereinafter Lott). Claim 16 The examiner respectfully points out that the combination with Puttagunta teaches generating, with an interpolation component, supplemental synthetic data using an interpolation algorithm (Puttagunta, [0005] — “Aspects may include a method of augmenting camera devices with neural networks to control framerate, time synchronization, image stitching and resolution, the method comprising creating synthetic camera frames by using neural networks to interpolate between actual frame captures”; that the interpolation is being performed discloses the existence of an interpolation component) and the synthetic data becoming part of at least one of the CNN test dataset or the CNN validation dataset (the combination teaches using synthetic data in the CNN test and validations datasets). The combination with Puttagunta fails to teach combining the supplemental synthetic data with the synthetic dataset, whereby the supplemental synthetic data becomes part of at least one of the CNN test dataset or the CNN validation dataset. Lott teaches combining supplemental synthetic data with a synthetic dataset ([0062] — “Further, the comparisons between the datasets may include comparing the original synthetic dataset with the additional synthetic dataset, comparing the original synthetic dataset with a combined dataset comprising the original synthetic dataset and the additional synthetic dataset…”). It would have been obvious to a person having ordinary skill in the art, before the effective filing date of the claimed invention, to combine the supplemental synthetic data with the synthetic dataset as taught by Lott before partitioning it, thereby becoming part of the CNN test and validation datasets as taught by the combination, to reduce bias from the GAN in the synthetic dataset by including synthetic data made without using the GAN. Prior Art The prior art made of record and not relied upon is considered pertinent to the applicant’s disclosure: Lu et al. (US 20190032153 A1), Authentication of Botanical DNA Isolated from Dietary Supplements Beavers, Randy (US 20080038196 A1), Standardization of Botanical Products Utilizing Biological Activity as a Marker Yu et al., Rapid Detection of Adulteration in Powder of Ginger (Zingiber Officinale Roscoe) by FT-NIR Spectroscopy Combined with Chemometrics. Elsevier, Food Chemistry: X, Volume 15, 2022, 100450, ISSN 2590-1575, https://doi.org/10.1016/j.fochx.2022.100450 Han et al., A Sense of Ginger Fraud: Prevalence and Deconstruction of the China-European Union Supply Chain. NPJ Sci Food 6, 51 (2022). https://doi.org/10.1038/s41538-022-00166-y The prior art listed above was used by the examiner to contextualize the claimed invention within the current state of the art. Conclusion Any inquiry concerning this communication or earlier communications from the examiner should be directed to RYAN JAMES STEAR whose telephone number is (571)272-8334. The examiner can normally be reached 7:30-5:30 EST/EDT. 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, Arleen Vazquez can be reached at (571) 272-2619. 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. /RYAN JAMES STEAR/Examiner, Art Unit 2857 /Catherine T. Rastovski/Supervisory Primary Examiner, Art Unit 2857