Apply 2D Image Convolution

Problem

Given a 2D grayscale image image as a list of lists of integers and a square filter kernel as a list of lists of integers, implement a function that returns a new 2D image produced by applying the kernel to every pixel using convolution-like neighborhood summation. Assume zero-padding outside image boundaries, and clamp each output value to the range [0, 255].

Constraints

• 1 <= rows, cols <= 200
• kernel is square with odd size: 1 <= k <= 7
• 0 <= image[r][c] <= 255
• -10 <= kernel[i][j] <= 10