Face image clustering method based on golden section method

Abstract

The invention discloses a face image clustering method based on a golden section method. The face image clustering method comprises the following steps: 1) applying a deep convolutional neural networkDCNN to realize feature representation of all face images in a database; 2) applying a K-Means ++ clustering algorithm to realize clustering of image representation; and 3) determining the optimal clustering number based on a 0.618 golden section method, and the process is as follows: firstly, giving a clustering range [a, b], and K belongs to [a, b]; randomly initializing a given clustering number K0 in the range, and constructing an optimization function f(K) based on an internal performance evaluation index of a clustering result; then, based on a 0.618 golden section optimization algorithm, dynamically searching the optimal solution of the function for in a one-dimensional mode; wherein the optimal solution is the optimal clustering number K*, and the corresponding clustering result C* is the optimal clustering of the face image library. According to the invention, the face image clustering performance is significantly improved.

Description

technical field [0001] The invention relates to a face image clustering method, in particular to a face image clustering method based on the golden section method. Background technique [0002] With the rapid development of computer vision and pattern recognition technology, image, as the most common visual information presentation mode, has broad application prospects. In the era of "big data", a large number of pictures are generated every day. For example, on social media, according to Facebook reports, an average of 350 million pictures are generated every day, most of which are face images. In judicial investigations, there are still a huge number of pictures that urgently need to be identified and classified. In terms of social security maintenance and monitoring management, a large number of face images captured by cameras need to be authenticated and stored for comparison. However, these face images usually do not have identity tags, or the tags are missing. In t...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the poor performance of existing face image clustering methods and to significantly improve the face image clustering performance, the present invention provides a face image clustering method based on the golden section method

Images