Method for filtering detection results of human skeleton key points under blurred background

• Main Authors: LI MENG, FANG LU, DAI QIONGHAI, WANG HAOQIAN, WANG XINGZHENG
• Format: Patent
• Language: Chinese; English
• Published: 07.09.2018
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The present invention discloses a method for filtering detection results of human skeleton key points under the blurred background. The method comprises: performing preliminary detection of human skeleton key points on an RGB image to obtain a preliminary result; performing depth estimation on the RGB image to obtain a depth estimation map; performing depth clustering on the depth estimation map to divide the depth estimation map into N regions, and calculating the average depth of each region, wherein N is greater than or equal to 1; and filtering the preliminary result, including, if N=1, not performing filtering, and taking the preliminary result as a final detection result, and if N is greater than or equal to 2, for each person who is subjected to the preliminary detection, respectively collecting statistics of the number the skeleton key points in the N regions of the depth estimation map, finding out the region with the largest number of skeleton key points for each person to determine the region to wh