opencv 学习笔记-入门（23）hog的快速积分图像-论文笔记
《An HOG-LBP Human Detector with Partial Occlusion Handling》阅读笔记：
It can filter out noises using the concept of uniform pattern. We believe that the appearance of a human can be better captured if we combine both the edge/local shape information and the texture information.
Trilinear interpolation and Gaussian weighting are two important sub-procedures in HOG construction.
In order to take the advantage of the integral image(积分图像法) without impairing the performance, we propose an approach, named as Convoluted Trilinear Interpolation (CTI), to do the trilinear interpolation. if we treat the feature value at each pixel as a 2D vector, we won’t be able to do the trilinear interpolation between pixels. To conquer this difficulty, we treat the feature value at each pixel as a 9D vector,
插值过程：
We designed a 7 by 7 convolution kernel to implement the fast trilinear interpolation. The weights are distributed to the neighborhood linearly according to the distances.
First, we need to vote the gradient with a real-valued direction between 0 and π into the 9 discrete bins according to its direction and magnitude. Using bilinear interpolation, we distribute the magnitude of the gradient into two adjacent bins(as shown in the top-right subplot of Figure 4). （将一个双线插值到两个毗邻的方向）
Then, the kernel in Equation ( 1) is used to convolve over the orientation bin image to achieve the trilinear interpolation.(用上面的kernel就可快速实现三线插值) The intermediate results are the trilinearly interpolated gradient image (bottom-left subplot of Figure 4), ready for integral image construction. （接下来进行积分的过程）
The intermediate trilinear interpolated results can be stored using the space allocated for the integral image. The trilinear interpolated gradient histogram image is of the same size as the integral image.（三线插值为积分图像分配出来。）
For each image, it is only a convolution with a 7 × 7 kernel, which can be further accelerated by Fast Fourier Transform (FFT).