Research, The Computer Automatically Puzzles Based On Color And Texture Features

The aim of Automatic puzzles solver is to automatically assemble an image of the pieces of a jigsaw puzzle by the shape, color and texture features of the pieces, Jigsaw puzzle assembly problem is a significant in the fields of machine vision, image analysis and pattern recognition. It contains a number of problems endemic to machine vision, like shape description, partial boundary matching, feature extraction, etc. It can be applied to diverse areas such as restoration of archaeological findings, repair of broken objects, biological research, etc.Most of the past research involving automatic puzzles and its related fields (restoration of archaeological findings, repair of broken objects) focus on the shape information contained on the individual pieces. They ignore the color and texture information contained on the individual pieces. However color and texture of the pieces is important features in the image processing field. This paper presents a new approach to the puzzle assembly problem that is based on using color features and texture features. A combination of the color and texture is used to match the puzzle pieces.The research contents of the paper include:1. Introduce the basic theory of the color and texture features in detail.2. An adjacent matching pieces retrieval algorithm based on color features is proposed. First the retrieved result in HSV color space is the best among others by the theory analysis and test results. Next in HSV color space, the color histogram of each pieces is extracted, which is compared with that of the retrieved piece. Absolute distance is the matching method in the algorithm. For a puzzle which contains 24(4x6) pieces, the number of pieces which right adjacent matching pieces in the top five in the array account of more than 75 percent of effective amount(20) by the algorithm. The retrieval result is satisfactory.3. An adjacent matching pieces retrieval algorithm based on texture features is proposed. The methods of texture features extraction is varied, The experiment results show that FFT is the best among the FFT and gray level co-occurrence matrix. For a puzzle which contains 24(4x6) pieces, the number of pieces which right adjacent matching pieces in the top five in the array account of more than 60 percent of effective amount(20) by the algorithm. The retrieval result is good.4. Combination of the color and texture is used to match the adjacent puzzle pieces. The adjacent matching pieces retrieval based on only one features is simple, Though the results is good, some pieces matching results is good based on color features but is bad based on texture features and vice versa. So the paper gives a new algorithm base on the color and texture. We analyze the performance based on the experiment results of nine puzzles, The results show that the number of pieces which right adjacent matching pieces in the top N in the array account of more than 73 percent of effective amount(N=effective amount×1/5, effective amount is the amount of pieces which have right adjacent piece) by the algorithm, The best results is 100%, So the algorithm is feasible, and have achieved good results.5. Introduce the implementation steps of adjacent matching pieces retrieval algorithm based on the combination of the color and texture features in detail.