Related Experiment Videos
Image recovery under nonlinear and non-Gaussian degradations
1Image Processing and Computer Vision Laboratory, Department of Electrical Engineering, Indian Institute of Technology Madras, Chennai 600 036, India.
Summary
A novel particle-filter-based recursive filter effectively recovers degraded images, even with complex nonlinearities and noise. This image restoration technique shows promise for various degradation types.
Area of Science:
- Digital Image Processing
- Computational Imaging
- Signal Processing
Background:
- Image degradation presents significant challenges in various applications.
- Existing image restoration methods often struggle with complex degradations like nonlinearity and non-Gaussian noise.
Purpose of the Study:
- To propose a new two-dimensional recursive filter for image recovery.
- To develop a particle-filter-based framework capable of handling general image degradations.
Main Methods:
- Utilized particle-filter theory to create a recursive image restoration framework.
- Obtained prior probability distribution samples via state model propagation.
- Computed sample weights based on measurement models and degraded images.
- Estimated the original image using the conditional mean of weighted samples.
Main Results:
- Demonstrated effectiveness in recovering images degraded by film-grain noise.
- Validated the method using both synthetic and real image examples.
- Showcased competitive performance compared to existing image restoration schemes.
Conclusions:
- The proposed particle-filter-based recursive filter offers a robust solution for image restoration.
- The framework shows potential for addressing a wide range of image degradations, including nonlinearity and non-Gaussian noise.