Local Features And Spatial Organization Of Them In Image Retrieval System

With the rapid development of multimedia technology and computer networking technology, the digital pictures and other multimedia data are increasingly available. The challenges for us in the digitizing epoch are how to place effective management on the abundant digital image information, as well as how to get what we need quickly from the large amount of information. Definitely, the technology of image retrieval comes into being to solve the problem. Recently, the content-based image retrieval (CBIR) has become a hot issue of the research. Many methods about CBIR have been proposed at present. However, the universal performance and accurate rate of retrieval are to be improved for most of these methods merely utilize the images'global properties.Local features, as an image feature extraction technology, have risen in recent years. Local features have proven to be very successful in applications such as image matching, image recognition, image mosaic, texture recognition, video data mining, recognition of object categories, image retrieval, and so on. Local features describe the local information of the images. Compared with the global features, the local features are more distinctive, invariable and robust. Thus, the local features are more adaptive when the image has blur background, partial occlusion or illumination changes. Using local features in the image retrieval system is very important for expressing the content of the image and matching the image features rapidly and precisely.The technology of local features contains two parts: local feature detector and local feature descriptor. And in the domain of image matching and image retrieval,the spatial geometry relationship of local features is also a important technology. This paper concludes the previous research and proposes related new methods to be aimed at the features of image retrieval. Finally, the new methods are applied to a content-based image retrieval system and achieve good results.The main contributions of this paper are:1. After the research and comparison on the existent local feature detectors, this paper proposed two improved methods of DOG detector. After considering every factor together we use the second improved method in our image retrieval system. And the experiment shows this method get good performance in the system.2. After the research and compare on the existent local feature descriptors, this paper studies the organizations of spatial relationship of local feature and uses this method in BOF algorithm.3. A content-based image retrieval system used in mobile phone is constructed based on the local features technology proposed by this paper and some related existent algorithm. We experiment on a corpus of image which has 100,000 images, and the test result shows the system gets good performance on the query precision and computes time.