Related Experiment Videos
A visual system model and a new distortion measure in the context of image processing
Journal of the Optical Society of America
|March 1, 1978
Summary
This study introduces a new image distortion measure based on visual system models, outperforming traditional squared-intensity differences in image restoration for better visual quality and detail. This advancement may also improve image storage and communication efficiency.
Area of Science:
- Computer Vision
- Image Processing
- Visual Perception
Background:
- Traditional image restoration relies on squared intensity differences, which are mathematically convenient but visually suboptimal.
- Existing distortion measures lack direct correspondence with human visual system models.
Purpose of the Study:
- To derive a new, visually relevant distortion measure from an accepted eye-brain system model.
- To compare the performance of this new measure against squared intensity differences in image restoration.
Main Methods:
- Developed an eye-brain system model comprising spatial frequency channels and image detectors.
- Derived a novel distortion criterion based on per-channel detection probability and phase angle changes.
- Applied optimal linear (Wiener) filters using both distortion measures to noisy images.
Main Results:
- The new distortion measure yielded superior image restoration compared to the squared-difference measure.
- Restored images were more visually agreeable, sharply detailed, and had truer contrast.
- The novel filter demonstrated impressive performance in image restoration tasks.
Conclusions:
- A new distortion measure derived from visual system models offers significant advantages for image restoration.
- This approach enhances visual quality, detail, and contrast in restored images.
- The mathematical properties of the new measure suggest potential for increased efficiency in image storage and communication.