Open ForawadStar opened 4 years ago
Sorry, perhaps I have not understood your question. Could you show your results of NMS? Thanks!
it is an example of NMS result, which was saved in output/nms/ folder after running eval_edge.m
I see. It looks strange since in my previous trial, each file should contain only one image, such as:
When I visualized the "mat" format image to be processed by non maximum suppression, I found that it did not encounter such a problem (each ". Mat" file contains only one image), but the result of NMS is that three images are concatenated along the column, so I think the problem may lie in the edgesNmsMex function or other functions. But such a problem is not fatal because I can extract one of the three NMS results as the final NMS.
Is it related to the version of MATLAB?My matlab is 2017b
Perhaps it is due to the MATLAB. I have only tried MATLAB R2018b (and some sub version of R2019) with the code in the repo. You may attempt to upgrade the MATLAB and generate the NMS results again.
Perhaps it is due to the MATLAB. I have only tried MATLAB R2018b (and some sub version of R2019) with the code in the repo. You may attempt to upgrade the MATLAB and generate the NMS results again.
ok, thank you
NMS requires a grayscale image with only one channel,while The edge image is read in RGB mode,leading to the three NMS result
Hmm that could one possible reason. Haven't tried it yet.
NMS requires a grayscale image with only one channel,while The edge image is read in RGB mode,leading to the three NMS result
You are right i think, this question that i have been paying attention.
I find that the results of NMS are the concatenation of three images along the column, is that correct?