This document summarizes techniques for feature extraction from video data to enable effective indexing and retrieval of video content. It discusses common approaches for segmenting video into shots and scenes, extracting key frames, and determining various visual features like color, texture, objects and motion. Feature extraction is an important but time-consuming step in content-based video retrieval. The document also reviews methods for video representation, mining patterns from video data, classifying video content, and generating semantic annotations to support search and retrieval of relevant videos.