Method for detecting model neck key points in e-commerce image

The invention relates to the technical field of image recognition, in particular to a method for detecting model neck key points in an e-commerce image, which comprises the following steps: step 1, determining the number Ns and positions Ls of human neck key points as labeling standards; step 2, labeling in the commodity image containing a whole body or the upper half body of the model; step 3, training a human neck key point detection model; step 4, inputting a to-be-detected image, and carrying out the detection; step 5, according to the detection result of the detection model, the number Npand the position Lp of detected model neck key points are given; and step 6, comparing the similarity between the matrix As and the matrix Ap. By means of the steps, the neck position of the model can be accurately detected, assistance is provided for further head posture recognition and commodity recognition of the image, it is unnecessary to prepare to manually classify commodities according tocollars, and labor cost i