A Semantic Image Segmentation Method Based on Region and Deep Residual Networks

Abstract

The invention discloses an image semantic segmentation method based on region and depth residual network. The region-based semantic segmentation method uses multiple scales to extract overlapping regions, which can identify targets of multiple scales and obtain fine object segmentation boundaries. Fully convolutional network-based methods use convolutional neural networks to learn features autonomously, which can be trained end-to-end for pixel-by-pixel classification tasks, but this method usually produces rough segmentation boundaries. The present invention combines the advantages of the two methods: first, the region generation network is used to generate candidate regions in the image, and then the image is extracted through a deep residual network with dilated convolution to obtain a feature map, and the candidate region and the feature map are combined to obtain The features of the region are mapped to each pixel in the region; finally, the global average pooling layer is used for pixel-by-pixel classification. The present invention also uses a method of multi-model fusion, setting different inputs in the same network model for training to obtain multiple models, and then performing feature fusion at the classification layer to obtain the final segmentation result. Experimental results on SIFT FLOW and PASCAL Context data sets show that the algorithm of the present invention has a higher average accuracy.

Description

technical field [0001] The invention belongs to the field of computer vision and relates to digital image preprocessing, model improvement, image semantic segmentation and simulation realization. Background technique [0002] Image semantic segmentation combines image segmentation and target recognition tasks. Its purpose is to divide the image into several groups of regions with specific semantic meanings, and mark the category of each region, so as to realize the reasoning process from the bottom layer to the high-level semantics, and finally obtain a A segmented image with pixel semantic annotation, that is, assigning a label to each pixel in the image representing its semantic target category. Image semantic segmentation has a wide range of applications in life, such as autonomous driving, geographic information systems, medical image analysis, and wearable application systems such as virtual or augmented reality. More and more emerging application fields require accura...

AI Technical Summary

Problems solved by technology

[0003] The region-based semantic segmentation method uses multiple scales to extract overlapping regions, which can identify targets of multiple scales and obtain fine object segmentation boundaries; the method based on full convolutional networks uses convolutional neural networks to learn features independently, which can target pixel by pixel Classification tasks are trained end-to-end, but this approach often produces rough segmentation boundaries

Images