Iterative bidirectional connection clustering algorithm for face image

Abstract

The invention discloses an iterative two-way connection clustering algorithm for face image recognition, which aims at the characteristic of high overlapping of face image data, adopts a data falling strategy on the basis of density clustering, finds out core data in an image data set, completes clustering work of the core data, and finally adopts a two-way connection mode to realize distribution operation of falling data. The method is characterized by comprising the following steps: calculating a distance matrix of a face image according to a formula; calculating shared neighbor similarity among the images; performing an iterative shedding process of the data to discover core data; core data clustering is completed; and completing the distribution of the remaining data through a bidirectional connection criterion. According to the iterative bidirectional connection clustering algorithm for face image recognition, each cluster is represented by a unique color, the outliers are represented by gray, the outliers are better judged, and more correct and more optimal cluster structures can be generated compared with an existing algorithm.

Description

technical field [0001] The invention relates to the field of image processing, in particular to an iterative bidirectional connection clustering algorithm for face image recognition. Background technique [0002] Face recognition is one of the research directions in the field of image recognition, which has very broad application prospects. Face recognition technology mainly includes three processes: face detection, face feature extraction and face recognition. However, with the increase in the "volume" of face databases and the increase in the requirements for "speed" of recognition, as well as the characteristics of non-uniform distribution of face image data points, traditional retrieval strategies are very time-consuming, which is not conducive to efficient recognition of face image features . [0003] Cluster analysis is an important method of data analysis. Its essence is to group data according to their characteristics, so that the similarity of data in the group is...

AI Technical Summary

Problems solved by technology

However, the clustering algorithms currently used in the field of face image recognition have the following problems: (1) the design of the density estimation function is complex, and when the data points of the face image set are not evenly distributed, it cannot be processed effectively; (2) Clustering is poor when the collection of images has a long and thin distribution

Images