Conventional methods for multimodal data retrieval use text-tag based or cross-modal approaches such as tag-image co-occurrence and canonical correlation analysis. Since there are ...
JungWoo Ha, Byoung-Hee Kim, Bado Lee, Byoung-Tak Z...
Robust local image features have been used successfully in robot localization and camera pose estimation; region tracking using affine warps is considered state of the art also for...
Biomedical images are invaluable in establishing diagnosis, acquiring technical skills, and implementing best practices in many areas of medicine. At present, images needed for in...
Daekeun You, Sameer Antani, Dina Demner-Fushman, M...
In using image analysis to assist a driver to avoid obstacles on the road, traditional approaches rely on various detectors designed to detect different types of objects. We propo...
Qi Wu, Wende Zhang, Tsuhan Chen, B. V. K. Vijaya K...
Conventional approaches to image retrieval are based on the assumption that relevant images are physically near the query image in some feature space. This is the basis of the clu...