Human body image hair detection method based on deep convolution neural network

Abstract

The invention discloses a human body image hair detection method based on a deep convolution neural network. The method mainly solves the problems of low detection efficiency and poor adaptability inthe prior art. The implementation scheme of the method comprises the following steps of 1) acquiring a super-pixel segmented images of a human body image; 2) marking the super-pixel segmented images;3) extracting image blocks from the human body image; 4) acquiring training samples from the image blocks; 5) constructing a deep convolution neural network and training the deep convolution neural network by means of training samples; 6) acquiring the super-pixel segmented images of the human body image at to-be-estimated hair positions; 7) extracting the image blocks of the human body image at to-be-estimated hair positions and classifying the image blocks by utilizing the well trained deep convolution neural network; 8) generating the hair detection result of the human body image at to-be-estimated hair positions. According to the invention, the complexity and the calculation amount of extracting characteristic operators are reduced. The robustness and the application range are improved. The method can be used for 3D printing, virtual fitting, human body testing and the construction of movie and television game models.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a method for detecting the position of hair in a human body image, which can be applied to 3D printing, virtual fitting, anthropometry, and video game model building. Background technique [0002] Object detection is a fundamental research topic in the field of computer vision. Its goal is to detect and recognize one or more known specific objects in an input image or to classify and localize all possible regions covered by predefined categories. Hair location estimation in human body images is an emerging branch in the field of object detection. [0003] Object detection is a prerequisite for a large number of advanced vision tasks, including activity or event recognition, scene content understanding, etc. And target detection has also been applied to many practical tasks, such as intelligent video surveillance, image retrieval, robot navigation, etc. Object det...

AI Technical Summary

Problems solved by technology

The shortcomings of this method are: First, because the algorithm needs to manually extract features from the image, when processing a large number of images, it is necessary to manually extract features every time, which greatly increases the complexity of the algorithm and reduces the efficiency. , secondly, due to manual extraction of image features, the most suitable features cannot be guaranteed to be extracted when processing human body images with complex backgrounds, which reduces the accuracy of the algorithm

The disadvantage of this method is that the number of hairstyles in the comparison database is limited. In practical applications, in the face of various hairstyles, the specific hairstyle of the image cannot be accurately determined, which has great limitations.

Images