Dear colleagues,
I have written the following MATLAB code, with two for...loops, which compute the distance map of an image. However, the code is relatively slower, and I want to speed it up using vectorization techniques. I am still a newbie in code optimization; still learning. Any idea on how to vectorize the code, please share:
Thank you Markus for the reply. I have noted the mistakes about row-column indexing you highlighted, and corrected accordingly. Also, I have tried the options and, for the second one, I get an error about "dimension mismatch". Troubleshooting the error, I printed sizes of matrices in the first for...loop, but they are all the same, and I wonder why matlab complains. While I still proceed troubleshooting, let me share with you the preliminary results. To run the code, you may set: v=ones(7,7); v(4,4)=0; d1=1; d2=1.(Please also note the comment lines 20 and 21, where I was trying to vectorize both rows and columns using the concept you provided. This, too, throws the same error. )
"bwdist" provides a similar function, however, according to the matlab documentation, this function is unoptimized for "Chessboard" and "City block" distances, which are my primary concerns. Besides, my interest was to implement the attached paper (page 344), but in a vectorized fashion. As you have put it, some present values of "v" depend on the previous values, and vectorization may thus be challenging.
- Badges
- Science topic
- Similar topics
- Vectorization