Page saliency element extraction method and system based on weighted hypergraph model

Abstract

The invention relates to a weighted hypergraph model-based page saliency element extraction method and system. The method comprises the following steps of: carrying out superpixel segmentation by utilizing a simple linear iterative clustering algorithm; constructing a common hypergraph model by using a fuzzy C-means clustering algorithm; according to priori knowledge of saliency object detection, performing weighting on vertexes and hyperedges in the common hypergraph model by using a position relation and color similarity between boundary superpixels and center superpixels to construct a weighted hypergraph model; and obtaining a transition probability matrix according to the transition probability matrix generation rule, and detecting elements in the page by combining the transition probability matrix and utilizing a random walk algorithm. According to the method, the weighted hypergraph model is constructed through the self-defined weighting strategy, the complex background in the image is weakened or removed, and only important elements in the page are extracted, so that the matching time cost is reduced, the method can serve an image matching algorithm more efficiently and accurately, and the image recognition efficiency and accuracy in the UI page automatic test are improved.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method and system for extracting salient elements of a page based on a weighted hypergraph model. Background technique [0002] Since the concept of salient object detection was proposed in 1998, there have been a large number of salient object detection algorithms based on traditional models and deep learning models. In the traditional salient object detection algorithm based on image features, the algorithm is improved in the two spatial dimensions of the air domain and the frequency domain; while in the algorithm based on the deep learning model, a large number of algorithms are pursuing the balance between time and precision. . [0003] When the existing image recognition algorithm is used for automated testing of UI pages, the automated testing efficiency is low and the test results are inaccurate due to the complex background of the image to be tested. Existing...

AI Technical Summary

Problems solved by technology

[0003] When the existing image recognition algorithm is used for automated testing of UI pages, the automated testing efficiency is low and the test results are inaccurate due to the complex background of the image under test.

Existing salient object detection algorithms can only obtain high precision saliency maps when the background of the detected image is simple or the contrast between the foreground and the background is obvious, and it still exists when it works on images with complex backgrounds or low contrast between the foreground and the background. Problems such as incomplete detection results, unclear boundaries of detected objects, etc.

Images