Multi-feature-combined Hash information retrieval method

Abstract

The invention relates to a multi-feature-combined Hash information retrieval method. The method is characterized by comprising the following basic steps that 1, an objective function is set up, data distribution of an object space is protected, compact matrix basis in an NMF is obtained, and redundancy is reduced; 2, alternative optimization is carried out, U and V are optimized through an iterative process, and updating rules of the base operator U and low-dimension data V are obtained; 3, global convergence is carried out, and alternating iteration is carried out through the original objective function; 4, a Hash function is generated, and finial results are obtained by calculating the hamming distance between training data and a test sample, namely XOR operation; 5, complexity analysis is carried out on the methods in the step 1 to step 4. By means of the method, probability distribution of data can be effectively protected, redundancy of low-dimension data is reduced, and therefore a Hash embedded function can be learned, wherein through the Hash embedded function, multiple expressions obtained from multiple sources can be fused, and RKNMF can be used for protecting high-dimension joint distribution and obtaining orthogonal basis.

Description

technical field [0001] The invention belongs to the technical field of computer information data processing, and in particular relates to a multi-feature joint hash information retrieval method for computer vision, data mining, machine learning or similar search. Background technique [0002] The learning of hash codes plays a key role in areas such as information processing and analysis, such as object recognition, image retrieval, and document understanding. With the advancement of computer technology and the development of the World Wide Web, large amounts of digital data require scalable retrieval of similar information. The most basic and essential method of similarity search is the nearest neighbor search: given a query image, find the most similar image in a huge database and paste the query image with the same image as the nearest neighbor. Tag of. Due to the large database in practical applications, nearest neighbor search is not a scalable linear search method (O...

AI Technical Summary

Problems solved by technology

Although the NMF method is superior to other existing methods, the existing NMF algorithm cannot solve the problem of protecting the local and overall structure of the original high-dimensional data, so there is a problem that the obtained low-dimensional data cannot inherit the high-dimensional data to the greatest extent.

[0010] To sum up, the deficiencies of existing technologies can be summarized as follows: First, because visual operators often have hundreds or even thousands of dimensions, most vision-based tasks will suffer from the curse of dimensionality; second, the previous Ha The Greek methods are mainly focused on a single feature. In their framework, only one feature operator is used to learn the hash function; the third is for effective similarity search. Although some multi-feature hash methods have been proposed, But the hashing of these methods is sensitive to noise and has high computational complexity

Images