Improved Image Retrieval Technology Based on Singular Value Decomposition

Abstract

An image retrieval technology combining singular value decomposition, bicubic interpolation and deep learning is proposed. Since the accuracy of image retrieval is affected by the quality of the input image, images with distinct features can accurately match the target image. For image retrieval and feature training that require input images of different scales, traditional image scaling methods will degrade image features. This paper uses singular value decomposition (SVD) and bicubic interpolation algorithm to improve the quality of the zoomed image, so that the deep neural network (DNN) can extract more accurate features. The experimental results show that the proposed algorithm can improve the accuracy of various image retrieval of power grid companies