High-Speed Tracking with Kernelized Correlation Filters（二）

      首先看看get_features函数。

            首先判断是hog特征还是gray，分两种情况。

            如果是hog特征，调用fhog函数，返回x，并将矩阵x的第三维最后一个组数据删除（好奇fhog函数http://vision.ucsd.edu/~pdollar/toolbox/doc/index.html）。

            

 if features.hog,
         %HOG features, from Piotr's Toolbox
         x = double(fhog(single(im) / 255, cell_size, features.hog_orientations));
         x(:,:,end) = [];  %remove all-zeros channel ("truncation feature")
                           %将矩阵x的第三维最后一个组数据删除
     end

            如果是gray，则直接对其归一化处理。

             

if features.gray,
        %gray-level (scalar feature)
        x = double(im) / 255;

        x = x - mean(x(:));
    end

函数最后返回x

    下面分析下各个核的correlation。

        首先是gaussian核，计算跟CSK没什么两样：

        

 N = size(xf,1) * size(xf,2);
     xx = xf(:)' * xf(:) / N;  %squared norm of x
     yy = yf(:)' * yf(:) / N;  %squared norm of y

     %cross-correlation term in Fourier domain
     xyf = xf .* conj(yf);
     xy = sum(real(ifft2(xyf)), 3);  %to spatial domain

     %calculate gaussian response for all positions, then go back to the
     %Fourier domain
     kf = fft2(exp(-1 / sigma^2 * max(0, (xx + yy - 2 * xy) / numel(xf))));

        

   对于linear kernel

kf = sum(xf .* conj(yf), 3) / numel(xf);

    对于polynomial kernel

1      xyf = xf .* conj(yf);
     xy = sum(real(ifft2(xyf)), 3);  %to spatial domain

     %calculate polynomial response for all positions, then go back to the
     %Fourier domain
     kf = fft2((xy / numel(xf) + a) .^ b);