Image Segmentation Method Simulating Human Vision

Abstract

The invention discloses an image division method simulating human vision. The method comprises the following steps of: 1, performing saliency detection on a target image to obtain a pixel saliency degree map; 2, ordering saliency points in the pixel saliency degree map according to the saliency degree; 3, selecting the first N saliency points to be used as a gazing point, and using the gazing point as the center to form a local region with the maximum information entropy, wherein the local region forms a gazing region; 4, performing random sampling in the gazing region, and performing equivalent pixel random sampling on the outside of the gazing region; and 5, obtaining a classification model by using an extreme learning machine training strategy; classifying all pixels of the target image; using a pixel region divided into a positive sample as a first division result; repeating the step 3 to the step 5 to obtain a repeated division result; and recording the region when the division result is stable; repeating the step 2 to the step 5 until no region capable of being gazed exists in the image. The image division method simulating human vision has the advantage that the human vision is simulated through gazing point ordering and a neural network model, so that the division on the target image is realized.

Description

Technical field [0001] The present invention relates to the technical field of human vision simulation, in particular to an image segmentation method that simulates human vision. Background technique [0002] With the development of information technology, computer vision has been widely used in low-level feature detection and description, pattern recognition, artificial intelligence reasoning, and machine learning algorithms. However, traditional computer vision methods are usually task-driven, that is, many conditions need to be defined, and corresponding algorithms are designed according to actual tasks, which lack versatility; need to solve high-dimensional nonlinear feature spaces, large data volume, problem solving and real-time processing, etc. Problems make its research and application face huge challenges. [0003] The human visual system can work efficiently and reliably in different environments. It has the following advantages: attention mechanism, saliency detection, ...

AI Technical Summary

Problems solved by technology

However, when the mechanism of human visual perception is not yet fully understood, it is still difficult to construct machine vision with human visual characteristics. If a machine vision system that simulates human vision can be constructed, the saliency information Extraction and segmentation will inevitably have an important impact on the field of computer vision applications.

Images