数字识别M文件.docVIP

数字识别M文件 文件头: %----------------------------------------------------------------- %Digit_Recognition.m Developed by Rentian Huang, %Hope University,Distribute System Emailhope.ac.uk %----------------------------------------------------------------- clear all; p(1:256,1)=1; load digit net; %Load the BP Neural Networks which been trained test=input(Please input a digits image:,s); %Prompt for input digit image for recognize s=input(Please input the number of digits:); %Prompt for input number of digits x=imread(test,bmp); %read the input image xbw=im2bw(x,0.9); %Convert the image to Black and White image xbw=medfilt2(xbw);%Use medium filter if needed bw=xbw; result=; %The loops for image segmentation for n=0:s-1 p1=ones(16,16); %Create an 16*16 Black and whit image (all white) [c,l]=size(bw); %Get the size of input image 文件尾: late the column size different between image and 16 p1(i1+1:i1+i,j1+1:j1+j)=xbwn1{n+1}; %Convert the image to standard 16*16 image xp(1:16,n*16+1:(n+1)*16)=p1; %Save the number n digit image p1=-1.*p1+ones(16,16); %Reverse the value 0 and 1 with each other in the image for m=0:15 p(m*16+1:(m+1)*16,1)=p1(1:16,m+1); %Form the vector of input digit image to the BP NN end [a,Pf,Af] = sim(net,p); %BP NN simulate b=round(a);%The output digit from the network if b0 b=b+1; elseif b9; b=b-1; end result=strcat(result,num2str(b)); %Display the output of the NN end imshow(xp); %Display the input image title(The input Numbers image:) text(0,60,strcat(Recognition Result：,result),FontSize,10);%Display the results %-------------------------------------------------------------------------- 另一个 文件头: %qq:445282655 %所选样本：usps_train