CAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View Depth
arXiv:1904.02028
Abstract
Single-view depth estimation suffers from the problem that a network trained on images from one camera does not generalize to images taken with a different camera model. Thus, changing the camera model requires collecting an entirely new training dataset. In this work, we propose a new type of convolution that can take the camera parameters into account, thus allowing neural networks to learn calibration-aware patterns. Experiments confirm that this improves the generalization capabilities of depth prediction networks considerably, and clearly outperforms the state of the art when the train and test images are acquired with different cameras.
Camera ready version for CVPR 2019. Project page: http://webdiis.unizar.es/~jmfacil/camconvs/
References in corpus (11)
- Depth Map Prediction from a Single Image using a Multi-Scale Deep Network
- DeepVO: Towards End-to-End Visual Odometry with Deep Recurrent Convolutional Neural Networks
- DeMoN: Depth and Motion Network for Learning Monocular Stereo
- Joint 2D-3D-Semantic Data for Indoor Scene Understanding
- ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes
- PlaNet - Photo Geolocation with Convolutional Neural Networks
- Learning Depth from Single Images with Deep Neural Network Embedding Focal Length
- BA-Net: Dense Bundle Adjustment Network
- Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-Scale Convolutional Architecture
- Just-in-Time Reconstruction: Inpainting Sparse Maps using Single View Depth Predictors as Priors
- Deep attention-based classification network for robust depth prediction