Design And Implementation Of Distributed Image Retrieval System

With the launch of recommended models for vision perception in recent years, Content Based Image Retrieval(CBIR) has been applied well, and efficiency in image training and image retrieval is inevitably highly demanded. Therefore, as to the shortcomings which the traditional image training and retrieval methods have to face up when coping with mass images, more flexible image training and more accurate image retrieval approaches become a hot research problem.This thesis presents a Two-Level Index Based on Binary Substring(TLI) supporting the image incrementation training, by virtue of which a distributed image training and retrieval system is designed and finally implemented.The image training process first extracts the binary descriptors of image by means of BRISK algorithm and identifies them by unique natural number in ascending order, then hashes those binary descriptors respectively into where the distributed computing nodes are through particular resource mapping rules. This processing method not only contributes balancing the storage and computing load, also providing high ability in expansion. The highlight of this method lies in the suppuration of image incrementation training, which makes images in train Set be increased and deleted dynamically, while keeps the training outcomes of other images from being affected.The computational model Map Reduce is melted in to the process of image retrieval which has two reducing levels. One is the substring processing for binary descriptor, from which a candidate set of feature-similar descriptors would be obtained. The other is the image features point reducing through which process a set of similar images is got. Meanwhile, the reducing process is combined with image features vector storage system, therefore, computing nodes only need localized running and there exists no network communication among these nodes so as to improve the degree of parallelism of image computing.In the final testing part, a comparison testing is operated among TLI proposed in this article and current mainstream high-dimensional descriptor index algorithms such as Locality Sensitive Hash(LSH) and Vocabulary Tree(VT). In the following, a performance comparison is run between the distributed image retrieval system implemented in this article and the distributed vocabulary tree training system implemented in the laboratory project Mobile Augmented Reality, and the testing result tells that both TLT and VT are similar in the retrieval accuracy, which is five to ten percent higher than LSH’s. The retrieval time efficiency also goes closer and closer with the increasing of the computing nodes, however, the biggest merits of TLI lies in its suppuration of image incrementation training which enables it more applicable for the internet in that this applications scenario is known for its rapid growth of image sets.The innovative points of this article are shown as follows:1. It comes up with a Two-Level Index Based on Binary Substring(TLI) on the basis of the Binary Descriptor Substring Partition Theory in the literature [1], which resolves problems of both Hamming KNN query and Hamming Range query.2. On the strength of the Two-Level Index Based on Binary Substring(TLI), a distributed image training and retrieval system which could support image incrementation training is designed so as to deal with fast training of mass image sets and high-accuracy retrieval.