Object Recognition and Image Tagging : with SIFT features

Bok av MD Kamrul Hasan
Whereas textual data retrieval history dates back to the beginning of library systems, multimedia information retrieval is relatively a new idea. Image retrieval research started with the form of Content Based Image Retrieval(CBIR); however, the progress encountered a bottleneck due to the semantic gap between visual features and conceptual semantics. The research question is: How can a machine semantically describe a picture? Answering this question might be a major breakthrough in image retrieval research. The basic areas of Computer Vision research that correlate image semantics are object recognition and object categorization. Both these fields require a general solution for detecting arbitrary objects, and automatic image annotation task requires a generalized object recognition framework and a methodology for annotation generation. The goal of this research is to use aspects of object recognition technologies for automatic automatic image annotation.