Invention Grant
- Patent Title: System and method for blind image quality assessment
-
Application No.: US16443922Application Date: 2019-06-18
-
Publication No.: US10909409B2Publication Date: 2021-02-02
- Inventor: Ying Chu
- Applicant: SHENZHEN UNIVERSITY
- Applicant Address: CN Guangdong
- Assignee: SHENZHEN UNIVERSITY
- Current Assignee: SHENZHEN UNIVERSITY
- Current Assignee Address: CN Guangdong
- Agency: Metis IP LLC
- Priority: CN201510111983 20150313
- Main IPC: G06K9/48
- IPC: G06K9/48 ; G06T7/00 ; G06K9/62 ; G06K9/03 ; G06K9/52

Abstract:
The present disclosure relates to a system and method for codebook construction and use thereof in image quality assessment. A codebook including a plurality of code words may be provided. Each one of the plurality of code words may have a reference score relating to image quality. A test image may be acquired. One or more feature vectors may be extracted from the test image based on statistical independence between neighboring divisive normalization transform coefficients of the test image. A score may be generated based on a comparison between the extracted feature vectors of the test image and at least some of the plurality of code words in the codebook and the corresponding reference scores of the code words.
Public/Granted literature
- US20190392246A1 SYSTEM AND METHOD FOR BLIND IMAGE QUALITY ASSESSMENT Public/Granted day:2019-12-26
Information query