Some Algorithms on Preprocessing of Two-Tone Pictures for Image Analysis

Abstract

The basic aim of image processing is to extract the information under constraints like noise, geometric errror, blurring and limited resolution in pixel and gray levels. In many cases like fingerprints, handwritten and machine- printed texts, earth crack patterns etc, the pictures are bi-level in nature. In recognition and classification of these pictures, several operations like thinning and skeletonization, shrinking and elongation, edge or boundary detection etc. are necessary. The paper presents efficient algorithms for such operations. Results of execution of these algorithms on different bi-Ievel pictures have also been presented. The modification of the algorithms for operations on multi-level pictures have also been discussed in the paper.

Indexing Terms:

• Image Processing
• Thinning
• Edge Detection
• Shrinking and Propagation

Additional information

Notes on contributors

S.K. Parui

Parui, S K: Received the Bachelor of Statistics and Master of Statistics degrees from the Indian Statistical Institute in 1974 and 1975 respectively. He had a training in Computer Science in the Indian Statistical Institute in 1975–1976 and then joined the Business Efficiency Services, Calcutta as a Programmer in 1977. He is now a Senior Research Fellow in the Electronics and Communication Sciences Unit, Indian Statistical Institute. His research interests are in pattern recognition.

B.B. Chaudhuri

Chaudhuri, B B: Obtained his BTech and MTech in Radio-physics and Electronics from Calcutta University in 1972 and 1974. During 1975–77 worked at IIT Kanpur on Optical communication systems and joined in 1978 as Computer Engineer at Electronics and Communication Sciences Unit, Indian Statistical Institute. Shri Chaudhuri is at present. working on Pattern recognition and Image processing problems. A member of IEEE, Computer Society of India and Optical Society of India, he has twenty research publications to his credit.