Super-resolution image retrieval system based on deep hash

Abstract

The invention is applicable to the technical field of image retrieval systems, and provides a deep hash-based super-resolution image retrieval system, which comprises the following steps of: S1, performing super-resolution reconstruction processing on a to-be-queried image (a) and a similar image (a +), namely inputting the images into a VDSR layer; and S2, inputting images SR (a), SR (a +) and a-obtained by reconstruction of a and a + into a CNN layer to extract corresponding image features. According to the super-resolution image retrieval system based on deep Hash, before the image is input into the neural network for feature extraction, super-resolution reconstruction is firstly carried out on the image, the resolution of the image is improved to obtain more detailed features, then the more detailed features can be extracted, the retrieval precision is improved, and when back propagation is carried out, the retrieval efficiency is improved. Compared with automatic learning updating of hyper-parameters and weights of low-resolution images, the high-resolution images are more reasonable, and the network obtained through training in the mode can obtain a better result compared with an original network even if the low-resolution images are input.

Description

technical field [0001] The invention belongs to the technical field of image retrieval systems, in particular to a deep hash-based super-resolution image retrieval system. Background technique [0002] As the convolutional neural network shines in the field of images, image retrieval technology is becoming more and more mature. Image retrieval is a crucial technology in the fields of medical images and automatic driving. Therefore, how to improve the accuracy of image retrieval has become an important issue. Research hotspots at home and abroad. [0003] Image retrieval technology uses text descriptions to describe the characteristics of images, such as the author, age, genre, and size of paintings. Image retrieval technology that analyzes and retrieves image content semantics such as image color, texture, layout, etc., that is, content-based image retrieval technology. CBIR is a type of content-based retrieval (CBR for short). CBR also includes dynamic video , audio and o...

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a super-resolution image retrieval system based on deep hashing, aiming to solve the problem that the hash code of the query image of the image retrieval system is similar to the hash code of similar images, and is very different from the hash codes of dissimilar images , the problem that image features cannot be extracted more accurately

Images