Related Experiment Video
Updated: Jul 7, 2026
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
Automatic assessment of constraint sets in image restoration
1Dept. of Electr. Eng., Auburn Univ., AL.
Summary
This study introduces cross-validation to assess tentative information for image restoration constraints. A modified, computationally feasible method effectively retains helpful constraints while discarding harmful ones.
Area of Science:
- Image processing and computer vision
- Signal processing
- Machine learning
Background:
- Image restoration often uses a priori information via constraints.
- Information for constraints is frequently tentative and may be inaccurate.
- Assessing constraint validity before incorporation is crucial for effective image restoration.
Purpose of the Study:
- To introduce cross-validation as a method for assessing the validity of tentative constraints in image restoration.
- To develop a computationally feasible modification of the cross-validation procedure.
- To evaluate the performance of both full and modified cross-validation methods.
Main Methods:
- Cross-validation technique applied to evaluate constraint sets.
- Development of a modified cross-validation procedure for computational efficiency.
- Experimental evaluation of the proposed methods on image restoration tasks.
Main Results:
- Cross-validation effectively assesses the validity of tentative constraints.
- The modified cross-validation procedure offers a feasible implementation.
- Both full and modified procedures demonstrated excellent performance in experimental results.
Conclusions:
- Cross-validation is a valuable tool for selecting reliable constraints in image restoration.
- The modified cross-validation provides a practical approach without significant performance loss.
- This work enhances the reliability and efficiency of incorporating prior information in image restoration.
