[2D Convolution, 20 pts] We have a 5x5 2D image shown in the following, 5 0 3 1 1 2 2 3 im 13 0 6 ILO 5 3 15 20 11 00 3 18 4 8 4 IN 2 6 5 7 3x3 Kernel: 3 3 2 1 1 0 0 1 5x5 Kernel: 3 1 102 1 2 3 1 1 0 0 2 1 1 2 0 1 1 1 3 1 4 1 IN lo a. We have two kernels of different sizes 3x3 and 5x5. Compute the final results of convolving this image with 3x3 kernels twice with zero padding and the results of convolving this image with 5x5 once with zero padding. (Hints: you can either compute by hand or you can use APIs from scipy or matlab packages. No need to show the steps but if using APIs, need to give a screenshot. If using the API, make sure you want to rotate by 180 degrees of the kernel beforehand.) b. Here, computational complexity is dominated by the number of multiplication operations. Count the number of multiplications in each of these operations: 1) convolve with 3x3 twice; 2) convolve with 5x5 once, and discuss which one is more computational-efficient