Integrated Search Based on Image Contents

Abstract

This works describes a new metaphor to make a contents search in an image database and to recover the results obtained using an automatically selected bandwidth. The increasing use of electronic commerce and the consequent publication of image catalogues through internet make evident the necessity to incorporate alternative mechanisms to the traditional search by keywords. Our work explores the scope of image-content access of databases. We use textual description, shapes, patterns and color similarities. At a later stage and due to the cost of the transmission of high quality images, we analyze the available bandwidth and we adapt the size of the resulting images to cope with the user requirements.