Authors:
Hongwei Zheng
and
Olaf Hellwich
Affiliation:
Computer Vision & Remote Sensing, Berlin University of Technology, Germany
Keyword(s):
Bayesian estimation, point spread function, blind image deconvolution, Mumford-Shah functional, partial differential equations, Γ-convergence, piecewise smooth approximate, graph-grouping, segmentation.
Related
Ontology
Subjects/Areas/Topics:
Applications
;
Artificial Intelligence
;
Biomedical Engineering
;
Biomedical Signal Processing
;
Computer Vision, Visualization and Computer Graphics
;
Data Manipulation
;
Enhancement and Restoration
;
Health Engineering and Technology Applications
;
Human-Computer Interaction
;
Image and Video Analysis
;
Image Filtering
;
Image Formation and Preprocessing
;
Image Quality
;
Methodologies and Methods
;
Model-Based Object Tracking in Image Sequences
;
Motion, Tracking and Stereo Vision
;
Neurocomputing
;
Neurotechnology, Electronics and Informatics
;
Pattern Recognition
;
Physiological Computing Systems
;
Segment Cluster Tracking
;
Segmentation and Grouping
;
Sensor Networks
;
Soft Computing
;
Software Engineering
;
Statistical Approach
;
Tracking of People and Surveillance
;
Video Analysis
Abstract:
We study a regularized Mumford-Shah functional in the context of joint prior models for blur identification, blind image deconvolution and segmentation. For the ill-posed regularization problem, it is hard to find a good initial value for ensuring the soundness of the convergent value. A newly introduced prior solution space of point spread functions in a double regularized Bayesian estimation can satisfy such demands. The Mumford-Shah functional is formulated using Γ-convergence approximation and is minimized by projecting iterations onto an alternating minimization within Neumann conditions. The pre-estimated priors support the Mumford-Shah functional to decrease of the complexity of computation and improve the restoration results simultaneously. Moreover, segmentation of blurred objects is more difficult. A graph-theoretic approach is used to group edges which driven from the Mumford-Shah functional. Blurred objects with lower gradients and objects with stronger gradients are grou
ped separately. Numerical experiments show that the proposed algorithm is robust and efficiency in that it can handle images that are formed in different environments with different types and amounts of blur and noise.
(More)