Image segmentation method, device and equipment and storage medium

Abstract

The invention discloses an image segmentation method, device and equipment and a storage medium. The method comprises the steps of obtaining an original picture; and inputting the original picture into an image segmentation model to obtain an image semantic segmentation image of the original picture, and generating the image segmentation model based on joint training of a superpixel segmentation algorithm and a full convolutional neural network. According to the embodiment of the invention, when image semantic segmentation is carried out on an original image by using an image segmentation model generated by training a super-pixel segmentation algorithm and a full convolutional neural network, the low-level features of the original picture are identified; the generated image semantic segmentation image is accurate in edge positioning and good in space consistency; the super-pixel segmentation algorithm only participates in the training generation process of the image segmentation modelbut not participates in the process of generating the image semantic segmentation graph by adopting the image segmentation model, so that compared with the image segmentation model generated only based on full convolutional neural network training, the prediction precision of the model is improved on the basis of not increasing the calculation overhead.

Description

technical field [0001] Embodiments of the present invention relate to computer vision technology, and in particular to an image segmentation method, device, equipment and storage medium. Background technique [0002] In recent years, with the improvement of computer hardware performance and the emergence of large-scale image data, deep learning has been widely used in the field of computer vision. Among them, the fully convolutional neural network is a deep learning neural network structure with outstanding achievements in the field of computer vision. [0003] Image semantic segmentation is one of the three core research issues in computer vision, and it is also the most difficult issue. Image semantic segmentation is to classify each pixel in the picture according to its category, and finally obtain a segmented image containing semantic information, that is, to classify each pixel in the picture into a predefined specific category and background category. For image seman...

AI Technical Summary

Problems solved by technology

[0004] In the process of realizing the present invention, the inventor found that there are at least the following problems in the prior art: First, the edge of the picture cannot be precisely positioned

However, because the full convolutional neural network does not fully consider the correlation between pixels with similar underlying features, the image segmentation model based on the training of the full convolutional neural network cannot fully recognize objects belonging to the same category, that is, it lacks spatial consistency. sex

The above existing problems make the prediction accuracy of the image segmentation model based on the training of the fully convolutional neural network not high

Images