Bootstrapping uncertainty in image analysis
Using the bootstrap resampling technique, it is demonstrated that uncertainty in the estimates of parameters in image models directly affects uncertainty in the image restorations themselves. Two resampling algorithms are developed which quantify this parameter uncertainty and use the information to improve the image quality. Real and artificial data were employed in numerical examples, and some unexpected results were obtained for the cross-validation choice of smoothing parameter.
Bibliographic Reference: Paper presented: Proceedings of the XII Symposium on Computational Statistics, Barcelona (ES), August 26-30, 1996
Availability: Available from (1) as Paper EN 39834 ORA
Record Number: 199610736 / Last updated on: 1996-08-06
Original language: en
Available languages: en