Semantic Classification of Sports News Video Using Color and Motion Features

With the growing popularity of digitized sports videos in the internet, an automated method to analyze semantic classification and content-based retrieval is essential to facilitate them. However, it is extremely difficult to categorize and label scenes in any video. This thesis uses the visual features(color and motion) of sports news videos to propose a method to extract certain scenes and analyze the sports genre and shows the experimental results after categorizing 5 sports games(soccer, baseball, golf, basketball, and volleyball).