A method and system for locating human body parts based on fast multi-dimensional space clustering

Abstract

The invention discloses a method and system for locating human body parts based on rapid multi-dimensional space clustering, wherein the method includes: obtaining intermediate results of human body part positioning, and importing the intermediate results in the form of point sets; Divide the tree, each node in the divided tree corresponds to each point in the point set, and each node is marked as unprocessed; any node in the specified point set is marked as unprocessed, and the node is marked as unprocessed Clustering processing, clustering processing divides all nodes in the tree into multiple different cluster centers; after all nodes are marked as processed, output all cluster centers that have been found. The present invention can be applied to an embedded system with limited computing resources to perform high-accuracy fast computing processing and aggregate identified points of different human body parts, without relying on prior training and easy to achieve global optimality.

Description

technical field [0001] The present invention relates to the field of information technology, in particular, to a method and system for locating human body parts based on fast multi-dimensional space clustering. Background technique [0002] In the human body recognition technology based on computer learning, a possible intermediate result is to complete the judgment of the body part of the human body part pixels in the image, and then need to use the clustering algorithm to find the center point of each body part and complete the skeleton. Construct. [0003] Clustering refers to merging points with close distance or similarity together. Clustering algorithms have a wide range of applications, and applications of clustering algorithms can be found in computer vision, data mining, and search engines. Existing clustering algorithms include K-means algorithm, CURE algorithm, nearest neighbor algorithm, grid-based STING algorithm, machine learning-based artificial neural netwo...

AI Technical Summary

Problems solved by technology

[0006] In view of the clustering algorithms in the prior art, it is difficult to avoid the problems of pre-training, long calculation time, low accuracy, and falling into local optimum under limited computing resources. The positioning method and system can perform high-accuracy and fast calculation processing in a limited computing resource environment, do not rely on prior training and are easy to achieve the global optimum

Images