Laser & Optoelectronics Progress, Volume. 57, Issue 12, 121101(2020)
Non-Reference Image Quality Evaluation in Color Channel
Fig. 1. Five images of different distortion types. (a) Additive Gaussian noise; (b) additive noise in colorcomponent; (c) spatially correlated noise; (d) masked noise; (e) high frequency noise
Fig. 2. Five additive Gaussian noise images of different distortion levels. (a)-(e) Level 1-5
Fig. 3. MSCN coefficient distribution of 5 images with different distortion types. (a) HSV_H channel; (b) HSV_S channel; (c) HSV_V channel; (d) gray space
Fig. 4. MSCN coefficient distribution of 5 images with different distortion levels. (a) HSV_H channel; (b) HSV_S channel; (c) HSV_V channel; (d) gray space
Fig. 5. Flow chart of RGB_R channel image quality evaluation model
Fig. 6. Scores of the 11th type of distorted images for each color channel training model
Fig. 7. Scores of the 20th type of distorted images for each color channel training model
|
|
Get Citation
Copy Citation Text
Ziang Qiao, Tao Liu. Non-Reference Image Quality Evaluation in Color Channel[J]. Laser & Optoelectronics Progress, 2020, 57(12): 121101
Category: Imaging Systems
Received: Sep. 24, 2019
Accepted: Oct. 30, 2019
Published Online: Jun. 3, 2020
The Author Email: Qiao Ziang (shammgod@126.com), Liu Tao (opticmcu@cjlu.edu.cn)