This is a real-life example of a data structure (pixel sum) that we use all the time in various computer vision and graphics work. The code is highly performance critical: it is called millions of times per image, often with very large kernels (thousands of pixels). At the end of the text there is a header that shows the interface that needs to be implemented. In its constructor PS_create(), the class receives as its input an 8-bit two-dimensional pixel buffer (we tend to work with grayscale data or one channel at a time). At this stage the constructor can take its sweet time to set up its data structures. The input buffers are guaranteed to have <= 4096 x 4096 pixels. Any changes made into the input buffer later should not affect the results of future queries, i.e., the constructor should take its own copy of the input.

typedef struct PS {

/* Your data here */

} PS;

PS* PS_create (const unsigned char* s, int width, int height)

_________-

What is the data structure and the function PS_create that provides the above implementation? Must be written in C with no external libraries.