Sample auxiliary labeling method and device based on FairMOT and correlation filtering and electronic equipment

• Main Authors: BAI CHAOCHAO, SUN HAOHUI, BAO WENLONG, LI XINYU, LI WUSHEN, HUANG KUN
• Format: Patent
• Language: Chinese; English
• Published: 04.08.2023
• Subjects: CALCULATING; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention provides a FairMOT and correlation filtering-based sample auxiliary labeling method and apparatus, and an electronic device. The sample auxiliary labeling method comprises the steps of S1, obtaining an image set and a first frame image; s2, labeling processing and classification processing are carried out, and labeling targets and categories are obtained; s3, a KCF algorithm model and a FairMOT algorithm model are preset; s4, obtaining a first prediction labeling target; s5, obtaining a second prediction labeling target; s6, carrying out weight distribution; s7, calculating a confidence coefficient according to a weight distribution result, and selecting a target with high confidence coefficient as a prediction labeling target; and S8, repeating the steps S2-S7 for other frames of images except the first frame of image in the image set, and completing the auxiliary annotation of the sample video. According to the method and the device, more automatic content labeling is realized, the complexity