Research On Image Moments Based On Non-symmetry And Anti-packing Model

Image moment with translation,rotation and scaling invariance are widely used in image classification,pattern recognition and scene analysis,etc,so the study of moment calculation is of very important theoretical significance and application value.But the traditional moment calculation not only can have high complexity and time consuming on handling binary image,but also its value is not accurate.In recent years,many excellent algorithm are transferred from the study of binary image moment into gray image moment research,but they dose not achieve the aim of fast and accurate moment calculation.Due to the traditional technique of image block representation in binary image,the result of image block has many horizontal blocks and is not optimal.In this paper,we propose an image representation algorithm based on non-symmetry and anti-packing pattern representation modal.As NAM is a method of block scanning based on the principle of greedy and can achieve the approximate optimal results and greatly reduce the number of sub model,in the next step of processing is to speed the moment calculation.In order to be able to get the precise moment value,this paper uses the integral method which can be calculated in advance and consume little time.We use the binary plane decomposition technique to produce the gray and color image to many binary image and then use the NAM algorithm to calculate the moment on transferred from binary image to gray and color image moment.We can ignore the low bit plane to get the approximate moment because it has little information and low noise about image.In this paper,binary image moment algorithm of binary image and gray image compare with IBR and LBT algorithm,the experimental results show that the proposed algorithm can effectively reduce the number of block to achieve the aim of quick and accurate moment calculation.Gray image moment algorithm compare with a variety of classic algorithm,the experimental result show that the proposed algorithm are better than them.