Research And Application On A Model Of Handwritten Digit Recognition Based On DFS

This paper analyse the image enhancement, the binarization, the noise abatement, the thinning, the normalization and the feature extraction on the handwritten digit recognition.We try to resolve the dynamic fuzzy multi-dimensional problem on the handwritten digit recognition. The main work is done as following:1.A model of dynamic fuzzy multi-dimensional data is proposed. The compose and the categoricalness theorem of the model of dynamic fuzzy multi-dimensional data is proposed. The provement of both is proposed.2.The model of dynamic fuzzy multi-dimensional data was applied to resolving handwritten digit recognition problem. The primary confirmation system is presented.The experimental results validate the good performances.Through the work above, the achievement is that we propose the new method on the handwritten digit recognition, another aspect provides a method of solving the problems of pattern recognition with high-dimensional and dynamic fuzzy.