倾斜haar-like feature计算

%generate_RSAT.m %2013.11.16 %by: zhou dongao %Function: %Caculate rotate 45 degrees summed area table %This code is based on the article of Rainer Lienhart's algorithm %See the pdf: %http://www.multimedia-computing.de/mediawiki//images/5/52/MRL-TR-May02-revised-Dec02.pdf %All of the parameters' names are cited from Rainer's article %Attention: Please read the article first! It's very useful to understand %this code function [RSAT] = generate_RSAT( image ) %test % clear all, close all,clc % image = double(imread('n01-x.PNG')); % figure, imshow(image), title('Input image'); %RSAT: rotate 45 degree summed area table, which size equal to image's %image: input image image = double(image); [H,W] = size(image);%width(W) is column,height(H) is row of the input image %expand the image's size by comple makeup of 0, %a row and two column are increased which valued 0 RSAT_ex = double(zeros(H+2, W+2));%initialize expanded RSAT and named RSAT_ex image_ex = [zeros(1,W); image]; %image_ex = [zeros(H+1,2), image_ex]; for j = 3 : H+2 %from top to bottom for i = 2 : W+1 %from left to right RSAT_ex(j, i) = RSAT_ex(j-1, i-1) + RSAT_ex(j-1, i+1)... -RSAT_ex(j-2, i) + image_ex(j-2, i-1) + image_ex(j-1, i-1);%see algorthm of Rainer end end % for j = 3 : H+1 % for i = 2 : W+1 % RSAT_ex(i, j) = RSAT_ex(i-1, j-1) + RSAT_ex(i+1, j-1)... % -RSAT_ex(i, j-2) + image_ex(i, j-1) + image_ex(i, j); % end % end RSAT = RSAT_ex(3:H+2, 2:W+1);%excluding zeros to make it's size equal to image's