In this paper, we present a novel image representation that renders it possible to access natural scenes by local semantic description. Our work is motivated by the continuing effo...
The structural similarity (SSIM) index has been shown to be a good perceptual image quality predictor. In many real-world applications such as network visual communications, howev...
The two major problems raised by a region-based image retrieval system are the automatic definition and description of regions. In this paper we first present a technique of unsup...
Coding at the sample level in still image watermarking takes advantage of avoiding a non-optimum initial diversity stage, used in many watermarking systems for tailoring Gaussian ...
We introduce flexible algorithms that can automatically learn mappings from images to actions by interacting with their environment. They work by introducing an image classifier i...