Study on Recognition of CAPTCHA Based on Attribute Grid Computing
Feng Jia-li · Computer Technology and Development · 2013
A new method is expected for CAPTCHA recognition in this paper,which is based on research on attribute grid computing,and the experiments show this method is feasible.There are several steps in the preprocessing of CAPTCHA,which include transforming image to grayscale image,image binarization,image denoising,image segmentation and normalization.In recognition part,use attribute grid computer,which is a kind of new computer based on qualitative mapping,build a attribute grid for object with its attributes and compute according to it.In this paper,build a attribute grid for standard characters,then,once need to recognize character,this grid can be used for computing and getting the results.The success rates of experiments are high,which verified the feasibility of this method,and also veified that the success rates of recognition are related with the allocation of the weights among different attributes.