Explaining Deep Neural Networks for Point Clouds using Gradient-based Visualisations.
Jawad TayyubMuhammad SarmadNicolas SchönbornPublished in: CoRR (2022)
Keyphrases
- point cloud
- neural network
- structure from motion
- surface reconstruction
- point sets
- point cloud data
- multi view reconstruction
- iterative closest point
- lidar data
- laser scanner
- triangular mesh
- stereo camera
- normal vectors
- semantic description
- point correspondences
- icp algorithm
- dimensionality reduction
- multi view stereo
- object recognition
- computer vision
- machine learning