A new approach, based on maximum likelihood, is developed for binary object image restoration. This considers the image formation process as a stochastic process, with noise as a random variable. The likelihood function is constructed for the cases of Gaussian and Poisson noise. An uphill climb method is used to find the object, defined by its "grain" positions, through maximizing the likelihood function for grain positions. In addition, some a priori information regarding object size and contour of shapes is used. This is summarized as a "neighbouring point" rule. Some examples of computer generated images with different signal-to-noise ratios are used to show the ability of the algorithm. These cases include both Gaussian and Poisson noise. For noiseless and low noise Gaussian cases, a modified uphill climb method is used. The results show that binary objects are fairly well restored for all noise cases considered.
| Identifer | oai:union.ndltd.org:arizona.edu/oai:arizona.openrepository.com:10150/276574 |
| Date | January 1987 |
| Creators | Li, Ming De, 1937- |
| Contributors | Frieden, B. Roy |
| Publisher | The University of Arizona. |
| Source Sets | University of Arizona |
| Language | en_US |
| Detected Language | English |
| Type | text, Thesis-Reproduction (electronic) |
| Rights | Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author. |
Page generated in 0.0019 seconds