Overview

Problem Statement

Beef is one of the most common meats, and the grading system is mature. According to the U.S. Department of Agriculture (USDA), there are three available grades of beef, Prime beef, Choice beef, and Select beef. Grades are determined by the amount of marbling in the meat, which can be approximately quantized as the ratio of marbling.

However, the current method to identify and determine a steak’s quality is still being done manually, which is susceptible to human error [1]. Besides, many people don’t know how to grade a steak accurately or judge whether it is a good piece for cooking.

Our goal is to develop a system based on computer vision and neural networks to help people evaluate the grade of steaks.