No-reference image quality assessment method based on saliency strategy and feature fusion
A technology of reference image and quality evaluation, applied in image analysis, image enhancement, image data processing and other directions, can solve the problem of insufficient consistency of evaluation results, and achieve the effect of improving feature representation ability
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0026] The embodiment of the present invention provides a non-reference image quality evaluation method for multi-scale feature fusion based on saliency strategy, such as figure 1 As shown, the method includes the following steps:
[0027] 101: Saliency Filter-Based Image Preprocessing
[0028] The image to be evaluated is preprocessed, converted into a grayscale image and divided into small non-overlapping image blocks, and the saliency analysis algorithm (Graph-based Visual Saliency, GBVS) algorithm based on graph theory is used to calculate the saliency of each image block. sex score. All image blocks are sorted according to the significance score, and the 25% image blocks with the highest score are selected as input samples.
[0029] 102: Extract multi-scale features from processed image blocks
[0030] In the embodiment of the present invention, a convolutional neural network with a two-stream structure is constructed to perform multi-scale feature extraction on the in...
Embodiment 2
[0036] The scheme in embodiment 1 is further introduced below in conjunction with specific calculation formulas and examples, see the following description for details:
[0037] 201: Image Preprocessing Based on Saliency Filters
[0038] For each image, its grayscale image is computed, followed by its saliency matrix. Each pixel is assigned a value ranging from 0 to 255. A higher saliency value represents a salient image pixel. The grayscale image and the corresponding saliency matrix are sliced into image patches of size 32×32. Each image patch is given the same subjective score as the original image. Based on the saliency matrix, the saliency score of each image patch is expressed by the following formula:
[0039]
[0040] In Formula 1, S(m,n) is the saliency value of a pixel at position (m,n) in the i-th image block. M and N represent the size of the image block. The saliency score reflects the degree to which the image block attracts people's attention. The hi...
Embodiment 3
[0054] Below in conjunction with concrete experiment, the scheme in embodiment 1 and 2 is carried out feasibility verification, see the following description for details:
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


