search

XuyangBai / D3Feat

[TensorFlow] Official implementation of CVPR'20 oral paper - D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features https://arxiv.org/abs/2003.03164
MIT License
261 stars 38 forks source link

Training data generation #11

Closed BingCS closed 4 years ago

BingCS commented 4 years ago

Hi Xuyang,

Thanks for sharing your excellent work. I just want to learn some details about training data generation. For training point cloud fragments, did you select depth images in every 50 frames or fuse every 50 frames to generate ptclds?

Best , Bing

XuyangBai commented 4 years ago

Hi Bing,

Thanks for your interest. We follow the instruction of 3DMatch and fuse every 50 depth images to get one point cloud fragment for the training set. For the testing set, we just use the point clouds provided by 3DMatch.

Best, Xuyang.

BingCS commented 4 years ago

Many thanks!