A New Contour Tracing And Inner And Outer Boundary Labeling Algorithm In Eight-Connected Binary Images

Chain code representing binary image possesses structured representation and high ratio of data compression. It plays an essential role in moment calculation, geometric feature detection, document image analysis and character recognition. And it has been theoretically and experimentally proved that chain code provide a compact representation and preserve all the information of the images.The encoding efficiency for binary images to represent shapes of objects is crucial in image storage and transmission. Two kinds of boundary description algorithm are mainly used in binary image. One is based on run-length code which records each consecutive'run'and one of its applications is run-length coding. The other one is chain-code, which is widely adopted for representing digital curves in image analysis and processing.This paper first analyzes the two algorithms, the run-length code algorithm and all sides labeling algorithm which are both capable of preserving the connectivity and any other feature of the image. New contour tracing algorithm is given in detail about how it make full utilization of the automaton in the labeling method to avoid tracing contours more than once as well as the unique labeling mechanism while tracing the image simultaneously. The connectivity information is fully preserved and the inner and outer contours can be distinguished. Comparison of the performance between using labeling method and computing from run-length representation are shown in the following section. In experiments on six types of images of various sizes, we compare our method with the algorithm computing chain code from run-length code. The results show that our algorithm outperforms that in all six types of binary image in terms of computational speed especially when the size of the image grows bigger and more complex. It always meaningful to improve the encoding efficiency for binary images utilizing today's computing power.