VisualSEEk is a fully automated content-based image query system that enables users to search images by visual features and spatial arrangements. The system automatically extracts and indexes salient color regions from images, allowing for complex joint color/spatial queries. It integrates feature-based image indexing with spatial query methods, using color sets to represent regions and enabling efficient spatial queries. The system allows users to specify color, size, and spatial layout of regions, and it supports both absolute and relative spatial relationships. VisualSEEk decomposes images into regions with feature and spatial properties, enabling efficient comparison of images based on their regions. The system also supports automated region extraction, direct indexing of color features, and provides a flexible query interface. The system's performance is evaluated through various queries, including color/spatial queries, and it demonstrates improved image retrieval capabilities compared to non-spatial content-based approaches. The system is implemented in Java and includes components for user tools, query server, image/video server, archive, metadata database, and index files. It supports searching through a test-bed of 12,000 color images and is being extended to search over one million images and videos from the web. The system uses color sets and back-projection to extract color regions and computes color similarity using a quadratic distance metric. It also supports spatial queries, including absolute and relative locations, and special spatial relations such as adjacency, overlap, and encapsulation. The system's query strategy involves computing color set distances, spatial distances, and region sizes, and it efficiently processes queries by combining these factors. The system's performance is evaluated through experiments, showing that it outperforms traditional color histogram-based methods in retrieving image matches.VisualSEEk is a fully automated content-based image query system that enables users to search images by visual features and spatial arrangements. The system automatically extracts and indexes salient color regions from images, allowing for complex joint color/spatial queries. It integrates feature-based image indexing with spatial query methods, using color sets to represent regions and enabling efficient spatial queries. The system allows users to specify color, size, and spatial layout of regions, and it supports both absolute and relative spatial relationships. VisualSEEk decomposes images into regions with feature and spatial properties, enabling efficient comparison of images based on their regions. The system also supports automated region extraction, direct indexing of color features, and provides a flexible query interface. The system's performance is evaluated through various queries, including color/spatial queries, and it demonstrates improved image retrieval capabilities compared to non-spatial content-based approaches. The system is implemented in Java and includes components for user tools, query server, image/video server, archive, metadata database, and index files. It supports searching through a test-bed of 12,000 color images and is being extended to search over one million images and videos from the web. The system uses color sets and back-projection to extract color regions and computes color similarity using a quadratic distance metric. It also supports spatial queries, including absolute and relative locations, and special spatial relations such as adjacency, overlap, and encapsulation. The system's query strategy involves computing color set distances, spatial distances, and region sizes, and it efficiently processes queries by combining these factors. The system's performance is evaluated through experiments, showing that it outperforms traditional color histogram-based methods in retrieving image matches.