KojiKobayashi / iterative_closest_point_2d

Iterative Closest Point 2D with python and opencv
MIT License
30 stars 14 forks source link

the meaning of KNN parameter "responses" #4

Open Lichuchen opened 1 year ago

Lichuchen commented 1 year ago

Great work! I have some problems about the meaning of KNN parameter "responses" . It seams that this "responses" should be the label for every point of the target point set, so that if the shape of the point set is (10,2), then the "responses" should be (10,1),and the label should be the corresponding index of the point (eg. np.array([[0],[1],[2], [3],[4]])), why that in this the "responses" is a (10,) array? And in the ICP part, "T = cv2.estimateRigidTransform(dst[0, indeces], src[0, indeces], True)", why that both the dst and the src are index by the indeces, if we are trying to find the transformation matrix from one to the other? Thanks for your time~