In this paper, we adopt local Fourier transform as a texture representation scheme and derive eight characteristic maps for describing different aspects of co-occurrence relations...
The present paper proposes a new concept of image processing method based on vector representation for visual inspection test. The method was applied to detect defects and extract...
We have been developing new relevance feedback algorithms for Content-based Image Retrieval (CBIR) that allow the user to achieve more flexible query. In conjunction with the new...
Munehiro Nakazato, Charlie K. Dagli, Thomas S. Hua...
Abstract. The ability to learn from user interaction is an important asset for content-based image retrieval (CBIR) systems. Over short times scales, it enables the integration of ...
Today's Content-Based Image Retrieval (CBIR) techniques are based on the "k-nearest neighbors" (kNN) model. They retrieve images from a single neighborhood using lo...