Ask Your Question

dereyly's profile - activity

2013-10-23 09:44:34 -0500 commented answer Multi cascade detector. LBP

Thank you John.V. Is it means that each feature generate matrix with equal size of input image? If i have 64 features and image 640x480 then features matrix is 64x640x480 with local sum of rectangles. But we need binary test around each point then matrix is 8x64x640x480 binary image or comsum of binary images that uses to calculate histogram OR 64x640x480 of uchars [0 255], but i think we cant use cumsum on this representation. I not fully understand how fast LBP works. What final representation of data we needed (to calculate histograms of given rectengle in each point)?

2013-10-23 03:22:38 -0500 received badge  Supporter (source)
2013-10-22 09:40:41 -0500 asked a question Multi cascade detector. LBP

Hello. How to create multi cascade detector with one feature pool (shared features)? I have many detecotrs (LBP) different objects or veiws of object, but they uses own features. Second, I not understand what mean this LBP features: [63] x=0; y=1; width=2; height=4; In theory LBP is binary histogram. I have one idea that x y is offset and width height is sizes of one of 8 rectangles. (In theory there is squres whith sizes of all pateern 3 6 9 and etc). How can i interpret LBP features in OpenCV?