Neural Networks and Image Data Management

This paper describes the research prototype of a postrelational DBMS CHINOOK being implemented at the University of Colorado at Colorado Springs. CHINOOK is intended to manage ultra-large databases of digitized images and digitized one-dimensional data as well as text and tables. This paper discusses our neural network based approach to the image data management in CHINOOK. We report our initial results on neural network generated image signatures which are the basis of CHINOOK image search and retrieval mechanism. We expect the image signatures to support the retrieval of images by their content.