"Jos (10584)" wrote in message <email@example.com>... > "amira ali" wrote in message <firstname.lastname@example.org>... > > "Jos (10584)" wrote in message <email@example.com>... > > > "amira ali" wrote in message <firstname.lastname@example.org>... > > > > need help urgently > > > > > > > > i have location of interest points in matrix A(2*13 ) where 2 is x,y of point and 13 is total number of points for example: > > > > x y > > > > 72.6781959551728 46.7543453323974 > > > > 71.9155271184922 100.930596158669 > > > > > > > > and i have another matrix called flow (120*160) this matrix contain velocity of every pixel (v in x direction) and (v in y direction). all i need it how can i get the velocity of 13 interest points from this matrix flow so finally i must have > > > > x y v(x) v(y) > > > > 72. 46.7543 0.087396 -0.135287 > > > > > > Using x and y as indexes in a matrix requires them to be integers ... > > > > > > Your question is also not very clear. You have "another matrix called flow" which is 2D but you index a one-dimensional vector v, using v(x) and v(y). > > > > > > Please rephrase ... > > > > > > ~ Jos > > > > it's not one dimensional vector it will be 13 rows and 4 coulmns , i written one row as example (location of points and their velocities) > > > in matlab v(x) means "the value of v at position (=index) x". > > Perhaps you can give a very brief example of what you want to achieve with two points (= two x and y values) and a flow matrix which is, let's say, 3 by 4? It would be very helpful if you show the exact inputs and desired outputs. 1-i have img1 (120*160*3) and img2 (120*160*3) then i get optical flow between two images in matrix called flow (120*160*2) (matrix flow contain velocity of every pixel in x direction and y direction).
2- i apply sift detector the image represent flow form img1 to img2 (mean flow image) and i get location of set of interest points 13*2(13 point s and 2 is cordinates of point)
3- i want to get velocity of this points from matrix flow
simply that clear the meaning of optical flow which i read it
pixel at the coordinates p = (x; y) in an image I. If the pixel moves by u; v pixels between time t - 1 and time t, the brightness constancy assumption can be formulated as: I(p; t - 1) = I(x + u; y + v; t)