How can I construct a Neural Network in Matlab with matrix of features extracted from images? -
i have x matrix contains on every row features extracted images (one image 1 row) , y matrix has on every row classification criteria every feature vector/row in matrix x (y matrix 1 column). want feed these 2 matrices neural networks , train , want test new feature vector of image (new_ft) , able 1 of 2 classification criterias built in neural network model. want see group new feature vector belong to. tried training network in form: net = newpr(x,y,numhiddenneurons) , use [net,tr] = train(net,xx,y); outputs = sim(net,xx); error "dimensions of matrices being concatenated not consistent." on first line. matrices have sizes: x= 46 x 25750 double , y = 46 x 1 cell. tried giving y same size 46 x 25750, same problem.
could let me know problem here? thank you!
try if feature vector length of each image 1*59 , have 4 images ; feature vector length 4*59.and have classify in 2 classes.
so train data should (for matlab nntool class) 59*4 of vector length , train label of 2*4 vector length
train_label =
1 1 0 0 0 0 1 1
now use
[net,tr] = train(net,traindata,trainlabel);
it work
Comments
Post a Comment