A Step Towards Efficient Bayesian Signal Reconstruction
This paper presents a theoretical basis for a set of optimal filters for the reconstruction of piecewise-continuous one-dimensional signals, drawing from Bayesian networks and Kaiman filters. Results are presented for synthetic and real data, using both the optimal filters and a sub-optimal implementation. The results compare well with linear space invariant filtering or facet fitting approaches, and present a basis for the design of image restoration algorithms.
KeywordsCovariance Tate Decen
Unable to display preview. Download preview PDF.
- Judea Pearl. Probabilistic Reasoning in Intelligent Systems. Morgan Kaufmann Publishers, Inc., 1988.Google Scholar
- Arthur Gelb. Applied Optimal Estimation. MIT Press, 1974.Google Scholar
- J. W. Dickson. Image Structure and Model-Based Vision. PhD thesis, Oxford University, Department of Engineering Science, 1990.Google Scholar
- Paul B. Chou and Christopher M. Brown. The Theory and Practice of Bayesian Image Labeling, September 1988.Google Scholar
- M.J. Swain and L.E. Wixson. Efficient estimation for Markov random fields. In Image Understanding and Machine Vision ,1989.Google Scholar
- B.F. Buxton, H. Buxton, and A. Kashko. Optimization, Regularization and Simulated Annealing in Low-Level Computer Vision. In Ian Page, editor, Parallel Architectures and Computer Vision Workshop. OUP, 1987.Google Scholar
- J.F. Silverman and D.B. Cooper. Bayesian Clustering for Unsupervised Estimation of Surface and Texture Models. IEEE Trans. Pattern Analysis Machine Intell ,10(4), July 1988.Google Scholar
- J.F. Canny. Finding Edges and Lines. Technical Report 720, Massachusetts Inst. Technol., 1983.Google Scholar