Fast optimization depth hash image coding method and target image retrieval method

Abstract

The invention discloses a fast optimization depth hash image coding method and a target image retrieval method. The fast optimization depth hash image coding method comprises the steps: building a hash image coding model for a large-scale image data set based on a greedy strategy, and generating binary codes of all images through a depth hash coding network obtained after optimization; and when target image retrieval is carried out, rapidly obtaining similar images of the same kind of a query image by calculating the Hamming distance between the query image code and the database image code. According to the fast optimization depth hash image coding method, the problems of gradient disappearance and quantization errors are better solved in combination with the neural network, and the codingperformance is better; the training process of the deep network is completed with fewer iterations, and the training speed is higher; the fast optimization depth hash image coding method can be applied to various problems with discrete constraints, and the application range is wider; and the optimization speed of the deep neural network and the retrieval performance of the generated image code are further improved, and the retrieval precision of the large image database is effectively improved.

Description

technical field [0001] The invention belongs to the technical field of information retrieval, and relates to image processing and fast image retrieval technology, in particular to a greedy strategy-based rapid optimization depth hash image coding method and a target image retrieval method. Background technique [0002] With the advent of the era of big data, data in various fields is growing explosively. In such a large wave of data, how to retrieve the information you need has become an important and urgent research topic. Hash algorithm is an algorithm for quickly completing target image retrieval on large image data sets. As the feature representation of the image), the Hamming distance is obtained through the fast XOR operation between the binary codes, so that the approximate nearest neighbor image retrieval is completed after sorting (that is, the image that is most similar to the query image is found from the image database) . This representation of binary image fea...

AI Technical Summary

Problems solved by technology

Although the retrieval performance has been improved, HashNet and DSDH still have problems such as slow network training and limited application range

Images