How to calculate the intersection of two polygons?

Support Vector Machines: In practical, what does summing two kernels means, respectively computed on Foreground and Background of images?

• What does that mean in practice ? Is it logically true, for example, to compute histogram intersection kernels on two different regions, then sum these two resulting vectors element by element? Additional explanation from the paper : The paper i am implementing is treating foreground and background separately. They cropped images into foreground (let's call X) and background regions (let's call Y). The histograms to feed SVM is computed distinctly on two image regions. And, finally, they calculate the final kernel value as : xf, yf : foreground histograms xb, yb : background histograms ---- K(x,y) = Kf(xf,yf) + Kb(xb, yb) ---- where; K(x,y) : the final kernel value Kf(xf,yf) : the kernel for foreground Kb(xb,yb) : the kernel for background

• Answer:

  A proper answer would need a thorough reading of the paper. However, from your description, I can say that summing two kernels usually means that the distance between two images is the sum of the distances between their background histograms and their foreground histograms. Which means that you want the distance measure to penalize differences in either histograms equally.