num_qubits must be replaced with img_dimen

[SaashaJoshi]

While constructing a QuanvolutionalNeuralNetwork, I realized that to have restrictions on the max size of the kernel/filter and stride of the quanvolutional layer, information about the dimensions of the image is necessary.

This will perhaps bring a huge change to the pipeline since everything as of now (before this issue) depends on num_qubits = int(math.prod(img_dimen)).

[SaashaJoshi]

This change might be needed only for Quanvolutional structures (layer or neural network) for now.