Color is a widely used visual feature for content-based video retrieval. There are two main methods discussed in the document: block-based and global color feature extraction. The block-based method extracts color histograms from divided blocks of each video frame, while the global method extracts a single color histogram from the entire frame. These color features are then used to measure similarity between videos for retrieval. The document also discusses challenges with high-dimensional color histograms and methods to reduce dimensions like transforms and selecting significant colors. Overall the paper presents color-based video retrieval techniques and evaluates performance of the block-based and global methods.