convolution.py

def convolution(image, filt, bias, s=1):
    '''
    Confolves `filt` over `image` using stride `s`
    '''
    (n_f, n_c_f, f, _) = filt.shape # filter dimensions
    n_c, in_dim, _ = image.shape # image dimensions
    
    out_dim = int((in_dim - f)/s)+1 # calculate output dimensions
    
    # ensure that the filter dimensions match the dimensions of the input image
    assert n_c == n_c_f, "Dimensions of filter must match dimensions of input image"
    
    out = np.zeros((n_f,out_dim,out_dim)) # create the matrix to hold the values of the convolution operation
    
    # convolve each filter over the image
    for curr_f in range(n_f):
        curr_y = out_y = 0
        # move filter vertically across the image
        while curr_y + f <= in_dim:
            curr_x = out_x = 0
            # move filter horizontally across the image 
            while curr_x + f <= in_dim:
                # perform the convolution operation and add the bias
                out[curr_f, out_y, out_x] = np.sum(filt[curr_f] * image[:,curr_y:curr_y+f, curr_x:curr_x+f]) + bias[curr_f]
                curr_x += s
                out_x += 1
            curr_y += s
            out_y += 1
        
    return out

Shouldn't image and filt be 2D arrays?

What are n_c and n_c_f? Thanks