Is there a reason to specify the input shape twice.

[leyuan]

Please correct me if I am wrong, in the second hidden layer, you resize the input back to original size after max pulling it. Is there a particular reason for this?

Hand-Gesture-Recognizer/cnn_train.py

Line 24 in 4746b60

model.add(Convolution2D(32,3,3, input_shape=(200,200,1),activation='relu'))

Also, did you just train the network for 10 epochs to get around 90% accuracy? If the weights are randomly initialized, that's pretty amazing efficiency...