Ahn, Heejune

icon company Seoul National Univ. of Science and Technology, icon location Seoul, South Korea icon location Professor and PhD in EE.

Results 4 comments of Ahn, Heejune

your miOU implementation is not correct.

your mIOU code : meanIOU = sum(intersectCount)/sum(unionCount) is wrong. you should do : iu := numpy array for all classes meanIOU = np.nanmean(iu)

cuda and cudnn version

Thanks for your sharing the code. But, your code does not work on my machine with 3 Tian Xp GPUs. I spent some time and I found it run with...

Rotation Invariant?

Not necessarily. But you mean rotation insensitve to 15 degress level? The journal paper of scm mentions complete rotation invariant. Which can be achieved by setting the tangent at the...

Rotation Invariant?

Someone did like this. https://medium.com/machine-learning-world/shape-context-descriptor-and-fast-characters-recognition-c031eac726f9