GIST Feature Extraction for Image Libraries

This program is implemented in MATLAB and designed to extract GIST features from image libraries. The implementation includes test files that verify the program's functionality and evaluate its performance. The core objective is to achieve feature representation and similarity comparison for image libraries through GIST feature extraction. The algorithm employs spatial envelope representation to capture the holistic properties of scenes, using oriented Gabor filters at multiple scales to compute dominant spatial structures. Key functions include image preprocessing, filter bank application, and feature vector generation. This tool enables users to efficiently extract GIST features from image collections, facilitating various image processing and analysis tasks such as image retrieval, classification, and content-based organization. The implementation follows standard GIST descriptor methodology while providing configurable parameters for filter settings and feature dimensions.