Design of Moving Target Detection System Using Lightweight Deep Learning Model and Its Impact on the Development of Sports Industry

Table 2

Target tracking core area.

Core area	Regional representation
Selection criteria of candidate areas	The size and proportion of different areas may vary, so it is essential to search and filter all areas that can match when selecting memory areas. however, this method is time-consuming and laborious, and will produce many property management areas, so feature analysis is a vital link.
Feature representation of candidate regions	After the collective search and analysis of the region, the region’s overall candidate region can be obtained. The next step is to filter its characteristics. Given the diversity of factors in the region itself, the feature extraction will affect the final classification results. Therefore, in most cases, the haar-link features are used for processing.
Classification measures of candidate areas	Since different regions will produce different differences in the processing of different objects, the input features can be effectively classified after massive feature processing. support vector machines, decision trees and neural networks are commonly used. by calculating multiple features, a classifier with strong performance can be obtained.