STRICT: An Image Retrieval Platform for Queries Based on Regional Content

This paper presents a CBIR system that is based on a segmented representation of image content. It compares regional features using fuzzy similarity, which have been shown to be psychologically intuitive. We show that they can be aggregated to support four different type of original queries. The system also supports competitive queries to test different visual comparison measures, and lets the expert user manipulate the parameters of the functions involved.

• Type: Conference
• Year: 2004
• Where: CIVR
• Authors: Jean-François Omhover, Marcin Detyniecki
• Keywords: CIVR 2004, Fuzzy Similarity, Segmented Representation, Visual Comparison Measures