Trying to implement 2d image convolution with numpy throws ValueError: operands could not be broadcast together

Question

I am trying to implement my own algorithm for convoluting an image with a certain filter and getting a bit of help from this post. What I have so far:

My image in numpy with shape (510,510)

imageInArray = np.asarray(gray_image)
print(imageInArray.shape)

## it prints this
(512, 512)

My convolution filter:

conv_filter = np.array([[0,1,0],[1,-4,1],[0,1,0]])
print(conv_filter.shape)
##prints this:
(3, 3)

Then I am spliting the image in submatrices depending on the filter kernel size like this:

### split image in submatrices
def getsubmatrices(imgArray, partitioning_shape):
    view_shape = tuple(np.subtract(imgArray.shape, partitioning_shape) + 1) + partitioning_shape
    strides = imgArray.strides + imgArray.strides
    sub_matrices = np.lib.stride_tricks.as_strided(imgArray,view_shape,strides)    
    return sub_matrices

imgSubMatrices = getsubmatrices(imageInArray, conv_filter.shape)
print(imgSubMatrices.shape)
##prints an array of size (510, 510, 3, 3)

But then when I am using the einsum method of numpy to multiply the submatrices with the filters like this:

multiplied_subs = np.einsum('ij,ijkl>ijkl',conv_filter,imgSubMatrices)

I am getting the following error:

ValueError: operands could not be broadcast together with remapped shapes [original->remapped]: (3,3)->(3,3,newaxis,newaxis) (510,510,3,3)->(510,510,3,3)
    

Could someone help me understand why this doesnt work? since I have spliting my matrix is submatrices of 3x3 it shouldnt be a problem to use the einsum method, right? I really apprecite any help you can provide as I have been trying to understand whats wrong since a few hours