A Fast Method of Similarity Network Fusion for Data Clustering

Abstract

The present invention provides a fast method for similarity network fusion for data clustering, said method comprising: similarity network training and storage in advance, fast fusion of similarity network; wherein: said similarity network training and storage in advance refers to using The distance between samples calculates the similarity network of different features of the sample, and stores the network obtained in this process, which is used for the rapid fusion of the similarity network for new samples later. The present invention constructs similarity networks into different types of data similarity networks, effectively fuses these networks, and effectively and quickly predicts the category of new targets according to the fused network and partially calibrated samples; The processing effectively reduces the time complexity in online prediction, and at the same time, the accuracy rate suffers only a small loss; it can be applied to video to achieve online target classification with faster speed and higher accuracy rate.

Description

technical field [0001] The invention relates to a data fusion method, in particular to a fast method for similarity network fusion for data clustering. Background technique [0002] At present, intelligent surveillance video has attracted a lot of attention due to its wide range of applications. In intelligent video technology, automatic object detection and classification has become a main task of intelligent surveillance system. Surveillance video object classification has become a challenging task due to the complexity of motion, interlaced backgrounds, changes in object positions and viewing angles, etc. At the same time, real-time online target classification is often required in actual monitoring systems. [0003] Among the existing object classification methods, some methods use shape and motion information for classification. This type of method is relatively fast, but its robustness is also limited because the shape of the object in the video will vary with the vi...

AI Technical Summary

Problems solved by technology

[0005] Although the similarity network fusion method has many advantages, it is not suitable for the target online classification due to its slow classification speed for a single new sample.

Images