Research On Content-Based Image Retrieval In The Compressed Domain

Image retrieval technique plays a crucial role in multimedia application. Compressed domain based retrieval methods which benefit from the establishing and popularity of the image and video compression standards such as JPEG, MPEG, and JPEG2000 are recently proposed. They are burgeoning and promising technique because they extract feature vector based on the compression parameters and have a lower cost for computing and storing the indices.In the paper, some of exploratory research work has been done around the JPEG image on the content-based image retrieval techniques in the compressed domain. The main work can be summarized as follows:1. Some analysises are made about retrieval technique and the research status of content-based image retrieval in compressed domain, and made a full-scale discussion about features extraction, similarity measure, evaluation of retrieval capability etc.2. An introduction is made about the knowledge of CBIR in compressed domain, and the retrieval methods in the transform domain and spatial domain are fully discussed.3. A study is made about the technology of JPEG compresses images, and the methods of CBIR in compressed domain is discussed.4. A algorithm is proposed to extract the feature of dominant color of JPEG compresses images on incomplete decoding. The experimental results demonstrate the algorithm has good performance.