Target object matching method, system and device and machine readable medium

Abstract

The invention provides a target object matching method, a target object matching system, target object matching equipment and a machine readable medium. The target object matching method comprises thefollowing steps: obtaining one or more matching areas when one or more target objects are in different postures; affine transforming the one or more matching regions to corresponding reference regions in a reference object; and determining the matching degree of the target object and the reference object. The method is based on a bottom-up deep attitude method. Key points of all human bodies in awhole image can be directly detected, one or more human bodies are selected as target human bodies, then the key points in the postures of the target human bodies in continuous frame images are subjected to affine transformation to another human body, and the matching degree of the target human bodies and a reference human body is calculated according to the result obtained after affine transformation; according to the method, the consumed time is short, and the matching degree of the target human body and the reference human body can be calculated under the condition that multiple target human bodies exist.

Description

technical field [0001] The present invention relates to identification technology, and in particular, to a target object matching method, system, device and machine-readable medium. Background technique [0002] The traditional two-dimensional target object matching method relies more on the processing capability of the hardware, and requires additional sensor equipment to obtain the pose information of the target object, and then compare the pose with the reference target. Even methods that do not require additional equipment need to perform target object detection and target object pose key point positioning first, such as DeformablePart Model (DPM, a deformable component model, which is an algorithm for target detection), etc. These methods It is time-consuming and lacks real-time performance. The currently commonly used method is the detection of skeleton key points based on deep learning, but most of these depth methods are from top to bottom (Top-down: For the detecti...

AI Technical Summary

Problems solved by technology

For example, when the target object is a human body, the common method is to detect the human body first, and then locate the skeleton points of the human body. The combination of the two steps is often time-consuming

Images