Deep Learned Full-3D Object Completion from Single View
This work addresses 3D reconstruction for real-time applications like robotics and navigation, though it is incremental as it builds on existing methods with efficiency improvements.
The paper tackles the problem of 3D object completion from single depth views by proposing a deep convolutional neural network with an auto-encoder that learns geometric features offline, achieving 92.9% reconstruction accuracy at 30x30x30 resolution with only about 4 million weights, which is roughly one-fourth the size of the leading network.
3D geometry is a very informative cue when interacting with and navigating an environment. This writing proposes a new approach to 3D reconstruction and scene understanding, which implicitly learns 3D geometry from depth maps pairing a deep convolutional neural network architecture with an auto-encoder. A data set of synthetic depth views and voxelized 3D representations is built based on ModelNet, a large-scale collection of CAD models, to train networks. The proposed method offers a significant advantage over current, explicit reconstruction methods in that it learns key geometric features offline and makes use of those to predict the most probable reconstruction of an unseen object. The relatively small network, consisting of roughly 4 million weights, achieves a 92.9% reconstruction accuracy at a 30x30x30 resolution through the use of a pre-trained decompression layer. This is roughly 1/4 the weights of the current leading network. The fast execution time of the model makes it suitable for real-time applications.