This is my project model for the CSU 2024 Spring Hackathon with USDA
In my program meat can be scored from 100 to 1,100 points. I used an SSID to compare images with the images that had the average score (in points) of each meat grade ("select" grade that had 425.6630682 points, "choice" grade that had 535.7205937 points, "low" grade which had 501.3157895 points, and "Prime" which had 601.7 points). Each time the meat was compared with an image it was given a similarity score compared to the meat grade index image it was compared to (between 0, being unlike, and 1 being the same image). Using that meat can be 110 to 1,000 points, has a similarity index between 0 and 1, and fits into one of the four grades.
I know my program was biased because I used a different number of images to compare with each item but it may have worked well.