In case of depth image processing, noise removal is important. It may seem to be easy. But in my case, it was not. When I removed values more than 150, there was no noise. But in that case, low depth information was removed too. When I removed values than 180 or the upper, noise was not deleted perfectly. But I can extract depth value range exactly, I uses that way.
'Papers > Machine learning' 카테고리의 다른 글
| Depth Kernel Descriptors for Object Recognition (0) | 2015.05.13 |
|---|---|
| A Case Study of Mobile based Object recognition by BOF Model (0) | 2015.05.09 |
| image resizing (0) | 2015.04.23 |
| classification vs. identification (0) | 2015.03.14 |
| Backpropagation, Dropout and Simulated annealing (0) | 2015.02.06 |