Solving jigsaw puzzles by computer

This thesis presents an algorithm for automatic solving of jigsaw puzzles with the help of digital camera. Application uses web camera to capture a picture of unassembled pieces of the puzzle and as a result we get a picture of assembled jigsaw. The first part presents extraction of individual pieces that make up the puzzle and determining the types of all four sides and calculating the corner points for each piece. The characteristics of pieces are used in the second part, where comparison of pieces to each other is shown. Pieces are compared by shape as well as by color informations. Since comparisons are time demanding tasks, we introduce a jigsaw piece model, which helps to greatly reduce number of necessary comparisons and thus reduces execution time. In the last part, we show the process of merging pieces into an output image where we, with the results of comparisons, determine the connections between the pieces and assemble them one by one. Application was tested on several puzzles with different numbers of pieces at different resolutions of input image