Key point location method

Abstract

The invention relates to a key point location method. The key point location method includes the steps: inputting an image, utilizing a pre-trained full convolution network to obtain the response diagram of each key point of an object-of-interest in an image; based on the response diagram of each key point, utilizing a pre-trained point distribution model to obtain initial location of each key point; and based on a weighted constraint mean value drifting method, performing iteration adjusting on location of each key point, and finally obtaining the final location of each key point. The key point location method organically integrates the expression capability of data drive with the prior inference capability of model drive; the full convolution network used by the key point location method can effectively respond to rigidity and non-rigidity conversion of the object in the image; the point distribution model can effectively respond to the shielding situation in the image; and the weighted constraint mean value drifting can reasonably balance the effect of the full convolution network and the point distribution model, so that the robustness of key point location can be greatly improved.

Description

technical field [0001] The invention belongs to the technical fields of pattern recognition, computer vision, digital image processing, etc., and more specifically relates to a key point positioning method. Background technique [0002] Key point location is one of the important links in computer automatic processing of images. It aims to quickly and accurately locate the key points of objects of interest in the image with strong semantics, such as the corners of the eyes, nose and mouth in face images. [0003] As far as face key point positioning is concerned, in the existing technology, both the traditional model-driven method and the emerging data-driven method can better deal with near-frontal face pictures with little change in expression and slight occlusion. However, limited by the expressive ability of the model-driven method and the sensitivity of the data-driven method to outliers, the existing technology cannot handle face pictures with large head poses, exaggera...

AI Technical Summary

Problems solved by technology

However, limited by the expressive ability of the model-driven method and the sensitivity of the data-driven method to outliers, the existing technology cannot handle face images with large head postures, exaggerated expressions, and severe occlusions.

In addition, the existing technology usually initializes key point positioning based on the output of the face detector. This initialization strategy makes the positioning algorithm easy to fall into a local minimum, which may even cause the algorithm to fail completely in extreme cases.

Images