Real-time image semantic segmentation method based on lightweight convolutional neural network

Abstract

The invention discloses a real-time image semantic segmentation method based on a lightweight convolutional neural network. The method comprises the following steps: constructing a lightweight convolutional neural network; training the constructed lightweight convolutional neural network; and performing semantic segmentation on the image in the given scene by using the trained lightweight neural network. In the constructed convolutional neural network, a multi-path processing mechanism is fused, multi-spatial-scale features of pixels can be effectively coded, and the problem that multi-scale targets are difficult to distinguish is solved. Meanwhile, model parameters are greatly reduced in combination with deep-wise convolution, the constructed lightweight convolutional neural network onlyhas 90 million parameters which are far lower than those of an existing method, the purpose of model lightweight is achieved, and the real-time processing requirement is met. Besides, the lightweightconvolutional neural network is based on a full convolutional network, end-to-end training and reasoning are realized, and the training and deployment process of the model is greatly simplified.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a real-time image semantic segmentation method based on a lightweight convolutional neural network. Background technique [0002] The purpose of image semantic segmentation is to assign a semantic category label to each pixel in the image, which belongs to the pixel-level dense classification task. On the whole, semantic segmentation is one of the basic tasks paving the way for comprehensive scene understanding. More and more applications also acquire knowledge from image data, including autonomous driving, human-computer interaction, indoor navigation, image editing, enhanced reality and virtual reality etc. [0003] Image semantic segmentation methods can be divided into two categories: one is traditional methods, such as threshold-based segmentation, edge-based segmentation, region-based segmentation, graph theory-based segmentation, energy functional-based segmenta...

AI Technical Summary

Problems solved by technology

Various models based on FCN have significantly improved the accuracy of semantic segmentation, but such models usually have millions of model parameters, and the inference efficiency is low, which seriously hinders their practical application

Images