Optic flow computation

Develops a new parameter-estimation framework for the problem of recovering optic flow from time-varying imagery, used in the computer processing of images and motion. Essentially, the approach divides the information into conservation and neighborhood categories, recovers each as an estimate with a covariance matrix, then fuses the estimates on the basis of their covariance matrices to calculate the optic flow. No index. Annotation copyright Book News, Inc. Portland, Or.