Image Retrieval Algorithm Based On Machine Learning

With the development of the Internet and the popularization of various video hardware devices,the image data has exploded.How to efficiently retrieve large-scale images has become a hot topic in the field of computer vision.Traditional image retrieval algorithms can not effectively use massive image datasets,and the featureextraction capabilities are not enough,which greatly affects the retrieval performance.In recent years,the application of depth convolution neural network in the field of image retrieval has also made breakthrough progress.This paper mainly studies the large-scale image retrieval algorithm,and proposes a Hierarchical-hash algorithm based on depth neural network.The algorithm can extract both high-dimensional semantic features and a set of short hash codes simultaneously.The network structure is divided into two parts.The first part is the pre-feature extraction network based on deep convolution neural network,which can extract the high image Dimensional semantic features.The second part is a sparse self-coding hash network,which can extract a set of short hash codes on the basis of high-dimensional semantic features of images.This paper applies Hierarchical-hash network to large-scale image retrieval.and proposes a hierarchical retrieval algorithm based on Coarse-Fine search.This algorithm can improve the performance of large-scale image retrieval to a certain extent while ensuring retrieval efficiency.Finally,this paper implements the algorithm and compares its result with the results of other image retrieval algorithms,which proves the effectiveness and feasibility of the algorithm.