	US 12,106,334 B2
	Artificial intelligence-based system and method for grading collectible trading cards
Tina Anne Sebastian, Lewis Center, OH (US)
Filed by Tina Anne Sebastian, Lewis Center, OH (US)
Filed on Jul. 2, 2022, as Appl. No. 17/856,984.
Claims priority of provisional application 63/185,547, filed on May 7, 2021.
Prior Publication US 2022/0374947 A1, Nov. 24, 2022
Int. Cl. G06Q 30/00 (2023.01); G06Q 30/02 (2023.01); G06T 7/00 (2017.01); G06T 7/13 (2017.01)

CPC G06Q 30/0278 (2013.01) [G06T 7/0002 (2013.01); G06T 7/13 (2017.01); G06T 2207/30168 (2013.01)]

4 Claims

1. A method for grading a collectible trading card using a combination of image recognition algorithms and artificial intelligence on a predefined scale, the method implemented on a system comprising a processor and a memory, the method comprises:

receiving, an image of the collectible trading card from a user device;

converting the image to a grayscale image;

applying an edge detection algorithm to the grayscale image to extract edge features;

applying a threshold inversion algorithm to the grayscale image to extract contrast/centering features;

applying a wavelet transform algorithm to the grayscale image to extract texture/surface features;

applying a corner detection algorithm to the grayscale image to extract corner information;

applying a color filtering algorithm to the grayscale image to extract stain detection features;

applying an image sharpen algorithm to the grayscale image to obtain an output and comparing the output with the image to obtain out-of-focus information;

processing the edge features, contrast/centering features, texture/surface features, corner information, stain detection features, and out-of-focus information using a bag-of-visual-words model to obtain quantitative data; subjecting the quantitative data to a pre-trained machine learning model to obtain a grade for the collectible trading card, wherein the grade is associated to the collectible trading card;

training a machine learning model using a set of pre-graded training images of collectible trading cards to obtain the pre-trained machine learning model; converting each training image of the pre-graded training images to a grayscale training image;

processing each grayscale training image using a plurality of predefined algorithms to obtain a plurality of outputs, the plurality of predefined algorithms comprises the edge detection algorithm, the threshold inversion algorithm, the wavelet transform algorithm, the corner detection algorithm, the color filtering algorithm, and the image sharpen algorithm;

subjecting the plurality of outputs to the bag-of-visual-words model to obtain respective quantitative data for each output;

generating a feature vector for each output using the respective quantitative data, wherein the feature vector is configured to grade the respective output based on the respective quantitative data.