A Probabilistic Hypergraph Construction Method Based on Spatial, Color and Central Bias Priors

Abstract

The present invention provides a probabilistic hypergraph construction method based on space, color and central bias prior, including: over-segmenting the input image into multiple image regions, calculating the features of each image region, and defining each image region as The vertices of the probabilistic hypergraph; construct a spatial hyperedge for each vertex based on the spatial prior, construct a color hyperedge for each vertex based on the color prior, and construct a hyperedge for each vertex on the edge of the image based on the central bias prior The central bias hyperedge, the probability of each vertex in the hyperedge belonging to the hyperedge is equal to the feature similarity between the vertex and the centroid point; the hyperedge in the hypergraph is a collection of three hyperedges, and the hyperedge weight is defined as each The sum of squares of the probabilities that a vertex belongs to a hyperedge. The present invention fully considers the space prior, color prior and central bias prior to construct a probability hypergraph, which can effectively describe the complex relationship between image regions in complex natural scene images, and is helpful to perform significant optimization in complex natural scene images. Target Detection.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a method for constructing a probability hypergraph based on space, color and center bias priors. Background technique [0002] In recent years, some researchers have proposed graph-based methods to process images. These methods describe the input image with a simple graph describing the binary relationship between two vertices. In a simple graph, the image region is defined as the vertices of the graph, two similar image regions are connected as the edges of the graph, and the edge weight is defined as the similarity between the vertices. The simple graph can describe the information contained in the image concisely and conveniently, and has achieved certain results in the field of image processing. [0003] However, using a simple graph to describe the binary relationship between image regions cannot describe all the key information in the input image. In...

AI Technical Summary

Problems solved by technology

However, there are few existing studies on hypergraphs, and the key information contained in complex natural scene images cannot be described yet.

Images