The Study On Algorithms For Generation Of Tree Structure Of Contours And Image Edit Based On Region Object

In the field of digit image processing, chain code is one of the most important tools by tracking and recording the information of image boundaries, thus the whole information of image can be formed as a relatively concise record, and the complexity is decreased with less storage space.In the paper, a simplified automaton model was proposed based on boundary tracking automaton, the current model can simplify transition of the automaton's states and the input-output process, and it can also decrease the probability of making mistakes in engineering practice by lowing the difficulty of understanding and implementation of the edge tracking automata.Using the simplified automaton model, an optimized contour tree generation algorithm is presented. Compared to the current algorithm, it reduced the space complexity greatly; it has local process characteristics at the same time, and brought great convenience for the local manipulation, such as pick-up of the objects in the image, and text recognition, etc.In order to restore the binary images which are coded by the contour tree structure, using one marker color, close to the theoretical optimal time complexity, and a space complexity of less than half the total length of the contours, this paper proposed a filling algorithm to complete the multi-level nested region filling tasks in an exemplary manner.Base on representation of the text object in the text image by contour tree, the paper discusses the principles and implements ways of the object-based image editing system, and completed a relatively integrated implementation which greatly facilitates the text image editing and typesetting and opened up new ways of image editing.