Training and testing on noisy modelnet data

yanx27 / PointASNL

PointASNL: Robust Point Clouds Processing using Nonlocal Neural Networks with Adaptive Sampling （CVPR 2020）

MIT License

260 stars 34 forks source link

Training and testing on noisy modelnet data #29

Open rabbiahassan opened 4 years ago

rabbiahassan commented 4 years ago

Hy !

Thanks for sharing code. In your experiment with title "shape classification with noise" Did you train your netowork on actual (noise free) modelnet dataset and then tested it by replacing actual points by few noisy points ?