A Comparison of Noise Cleaning Techniques
Abstract
A variety of noise cleaning techniques are compared on two test pictures. All of the techniques tested are based on local analysis of a small neighborhood of each pixel. They include median and mode filtering; selective local averaging or weighted averaging schemes of various sorts; and Kalman filtering. Iterated weighted averaging and iterated median filtering were judged to give the best overall performance.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1978
- Accession Number
- ADA074854
Entities
People
- Judith P. Davenport
Organizations
- University of Maryland