Image Classification Method Based on Neighbor Supervised Discrete Discriminant Hashing

Abstract

The invention discloses an image classification method based on the nearest neighbor supervised discrete discriminant hash, obtains a standard image library, and constructs a new standard image library to be classified; calculates the inter-class weight matrix and the intra-class weight matrix of the new standard image to be classified, And get the intra-class hash relationship function and inter-class hash relationship function; select the anchor point for the collected image training samples in the standard image library to be classified, and calculate the distance between each training sample and the anchor point; get the final objective function; decompose the final objective function, and obtain the feature matrix; based on the feature matrix and use the nearest neighbor classifier to classify the image; the present invention can better describe the relationship between samples, and strengthen the reliability of local classes. Classification and intra-class compactness, so as to learn a more efficient hash function, and then learn a compact binary hash code, aiming to improve the accuracy of high-dimensional image recognition for massive data and speed up search and retrieval.

Description

technical field [0001] The invention relates to the technical field of image classification, in particular to an image classification method based on the nearest neighbor supervised discrete discriminant hash. Background technique [0002] The informatization of human production and living data has led to an explosive growth in the amount of data. Faced with these huge image libraries, people have proposed a series of feature extraction methods for data compression. The high-dimensional data processing methods proposed in the past There is room for improvement in terms of stickiness and scalability. The hash algorithm maps the original features to a new feature space by random permutation or projection, and converts it into a compact binary hash code. This not only improves the speed of image search and retrieval, but also improves the storage efficiency of large-scale data. Based on these advantages, the hash algorithm has received extensive attention in the fields of pat...

AI Technical Summary

Problems solved by technology

Most of the supervised hashing algorithms in the past only considered the label information of the training samples, but did not further consider the neighbor relationship between samples; therefore, it is necessary to design an image classification method based on neighbor supervised discrete discriminant hashing.

Images