Image Saliency Detection Method Based on Fusion of Texture Feature and Color Feature

Abstract

The invention discloses an image saliency detection method based on the fusion of texture features and color features, which mainly solves the problems of insufficient utilization of texture features and poor saliency detection effect on high-texture images in the prior art. The scheme is: 1) input the image, and use the filtering method based on the total variation model to remove the texture, and obtain an image containing color features; 2) filter the input image using a Gabor filter to obtain an image containing image texture information ; 3) Calculate the preliminary contrast value based on the color feature image; 4) Calculate the background probability based on the texture information image; 5) Fusion the preliminary contrast value and the background probability to obtain a new contrast, and then obtain a saliency map based on color and texture features. The invention makes full use of the color and texture information of the image, improves the detection effect of complex texture images, and can be used in computer vision tasks.

Description

technical field [0001] The invention belongs to the technical field of image detection, and in particular relates to an image saliency detection method, which can be used for image segmentation, target recognition, adaptive image compression, content-aware image scaling and image retrieval. Background technique [0002] People often effortlessly judge the importance of image regions and focus on the important parts. Since computing resources in image analysis can be allocated optimally through salient regions, it is of great significance to detect salient regions of images by computer. Extracting saliency maps is widely used in many computer vision applications, including image segmentation of objects of interest, object recognition, adaptive image compression, content-aware image scaling, and image retrieval. [0003] Saliency stems from visual uniqueness, unpredictability, scarcity, and singularity, and it is often attributed to variations in image properties such as colo...

AI Technical Summary

Problems solved by technology

[0006] All the methods mentioned above have a common problem: the texture information of the image is not considered, and the texture information is an important feature reflecting the intrinsic nature of the image

Castleman et al. believe that texture is the spatial distribution attribute of the gray level of pixels in an image. The inherent attribute of this spatial structure can be described by the correlation between neighboring pixels. The visual features that reflect the homogeneous phenomenon in the image reflect the intrinsic properties of the object surface. Due to the lack of texture features, the above methods will inevitably lead to a decline in the quality of high-textured images.

Images