A detection method and system for three-stage image quality assessment
An image quality assessment and quality assessment technology, applied in the field of intelligent recognition, can solve the problems of image quality deviation, poor image quality, error discarding, etc., to achieve high accuracy, strong interpretability, strong generalization ability and expression ability. Effect
Active Publication Date: 2022-07-15
ZHEJIANG UNIV +1
View PDF16 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
This kind of method has been proved to be effective to a certain extent, but this kind of design process currently considers the image quality of the whole image. In practical applications, the image quality of the concerned area is good, but the image quality of other parts of the area is poor, resulting in Erroneously discarded due to overall image quality deviation
[0004] Therefore, it is urgent to propose a detection method for image quality assessment to solve the problem of only evaluating the quality of the image area of interest
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0057] see figure 2 , collect images of Jiefang brand trucks from surveillance video frames, and perform three-stage image quality assessment detection. Target detection is carried out on the features of the video frames of Jiefang brand trucks, and the Jiefang brand trucks are taken as the target of attention, and their confidence scores, classification categories and target coordinates are obtained, and then screened. According to the classification category and target coordinates obtained by screening, the features of the video frame of Jiefang brand trucks are detected, and the "Jiefang" trademark, windshield and truck license plate number are used as components to obtain the confidence score, similarity score, component Coordinates and component features, take the license plate as a string attribute and the windshield as a category attribute, calculate the similarity with the specified attribute respectively, and filter according to the similarity. According to the conf...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a detection method and system for three-stage image quality assessment. The method models road monitoring pictures or videos, and outputs the category, location information and corresponding image quality of the object of interest. Specifically, in the first stage, the image or video frame is used as input, and the target feature and its position information of interest are output through the target detector; in the second stage, the target feature and its position output from the first stage are used. The information is selected and matched by the part detector to output the target part feature and position information. In the third stage, the feature and position information of the target component output in the second stage are passed through the image quality assessment classifier to output the quality level of the region of interest in the image. The method can accurately output the quality level of the corresponding image area based on the rapid detection of the object of interest, combined with the component information.
Description
technical field [0001] The invention belongs to the technical field of intelligent identification, and particularly relates to a detection method and system for three-stage image quality assessment. Background technique [0002] Image quality assessment has a wide range of applications in the industry, such as the screening of evidence of violations of laws and regulations in road traffic surveillance videos. The task of image quality assessment can be divided into three stages: defining goals, collecting human labels, and objective quality metrics for training data. Given an image or video frame, traditional image quality assessment methods select image regions that need attention and require a high technical threshold. With the rapid development of deep convolutional neural networks, image quality assessment methods based on deep convolutional neural networks have gradually increased. At the same time, the latest development of image quality assessment datasets with large...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/00G06V20/40G06V10/764G06V20/62G06K9/62
CPCG06T7/0002G06T2207/30168G06F18/24323
Inventor 卢朝晖齐国栋于慧敏王润发顾建波
Owner ZHEJIANG UNIV