Browse Prior Art Database

Chain Coding of Regions in a Multicolored Image

IP.com Disclosure Number: IPCOM000049800D
Original Publication Date: 1982-Jul-01
Included in the Prior Art Database: 2005-Feb-09
Document File: 3 page(s) / 61K

Publishing Venue

IBM

Related People

Danielsson, P: AUTHOR

Abstract

This invention relates to a method for extracting the regions of a non-binary image and converting these to a list of separate objects. The output data is not only a chain-coded contour for each object, but also for the adjacency conditions that represent the topology of the original image.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

Page 1 of 3

Chain Coding of Regions in a Multicolored Image

This invention relates to a method for extracting the regions of a non-binary image and converting these to a list of separate objects. The output data is not only a chain-coded contour for each object, but also for the adjacency conditions that represent the topology of the original image.

The steps of the method include (1) searching the image in row major order until an unlabeled border of a colored object is detached, (2) labeling the detected border edges and following the border to encounter successful pals (picture elements) of the same color value comprising the boundary of the object, and (3) resuming the row major order scan. Both sides of a border are followed and labeled in two different procedures which are called "island following" and "object following".

Morrin [1] disclosed the dual mode of raster scanning and border following of black-white images. Kruse [2] avoided the time-consuming "shrinking" of Morrin through the use of edge labeling and stack processing them. Kruse was not effective in the non-binary case. The method of this invention resolves contiguous differentially colored interior objects by looking for interior nonlabeled edges.

Referring now to Figs. 1. 2, and 3. there are shown flow diagrams of the method. Around the border of the full image the convention is adopted that there is a frame with the imaginary label 'frame' which is included in the new set NS = /A, B, C...,'frame'/. The variable m that is continuously updated to the present object Label during raster scan (box 8) is consequently set to 'frame' at the beginning of each line in box 2.

The first pixel encountered by a rast...