We present an unsupervised learning framework for the task of monocular depth and camera motion estimation from unstructured video sequences. In common with recent work [10, 14, 16], we use an end-to-end learning approach with view synthesis as the supervisory signal. In contrast to the previous work, our method is completely unsupervised, requiring only monocular video sequences for training. Our method uses single-view depth and multiview pose networks, with a loss based on warping nearby views to the target using the computed depth and pose. The networks are thus coupled by the loss during training, but can be applied independently at test time. Empirical evaluation on the KITTI dataset demonstrates the effectiveness of our approach: 1) monocular depth performs comparably with supervised methods that use either ground-truth pose or depth for training, and 2) pose estimation performs favorably compared to established SLAM systems under comparable input settings.

/pdf/unsupervised-learning-of-depth-and-ego-motion-from-video-1dcmjuaj7g.pdf

Unsupervised Learning of Depth and Ego-Motion from Video

We introduce Recurrent All-Pairs Field Transforms (RAFT), a new deep network architecture for optical flow. RAFT extracts per-pixel features, builds multi-scale 4D correlation volumes for all pairs of pixels, and iteratively updates a flow field through a recurrent unit that performs lookups on the correlation volumes. RAFT achieves state-of-the-art performance. On KITTI, RAFT achieves an F1-all error of 5.10%, a 16% error reduction from the best published result (6.10%). On Sintel (final pass), RAFT obtains an end-point-error of 2.855 pixels, a 30% error reduction from the best published result (4.098 pixels). In addition, RAFT has strong cross-dataset generalization as well as high efficiency in inference time, training speed, and parameter count. Code is available at https://github.com/princeton-vl/RAFT.

RAFT: Recurrent All-Pairs Field Transforms for Optical Flow

This paper presents a self-supervised framework for training interest point detectors and descriptors suitable for a large number of multiple-view geometry problems in computer vision. As opposed to patch-based neural networks, our fully-convolutional model operates on full-sized images and jointly computes pixel-level interest point locations and associated descriptors in one forward pass. We introduce Homographic Adaptation, a multi-scale, multi-homography approach for boosting interest point detection repeatability and performing cross-domain adaptation (e.g., synthetic-to-real). Our model, when trained on the MS-COCO generic image dataset using Homographic Adaptation, is able to repeatedly detect a much richer set of interest points than the initial pre-adapted deep model and any other traditional corner detector. The final system gives rise to state-of-the-art homography estimation results on HPatches when compared to LIFT, SIFT and ORB.

/pdf/superpoint-self-supervised-interest-point-detection-and-vny6abkges.pdf

SuperPoint: Self-Supervised Interest Point Detection and Description

Per-pixel ground-truth depth data is challenging to acquire at scale. To overcome this limitation, self-supervised learning has emerged as a promising alternative for training models to perform monocular depth estimation. In this paper, we propose a set of improvements, which together result in both quantitatively and qualitatively improved depth maps compared to competing self-supervised methods. Research on self-supervised monocular training usually explores increasingly complex architectures, loss functions, and image formation models, all of which have recently helped to close the gap with fully-supervised methods. We show that a surprisingly simple model, and associated design choices, lead to superior predictions. In particular, we propose (i) a minimum reprojection loss, designed to robustly handle occlusions, (ii) a full-resolution multi-scale sampling method that reduces visual artifacts, and (iii) an auto-masking loss to ignore training pixels that violate camera motion assumptions. We demonstrate the effectiveness of each component in isolation, and show high quality, state-of-the-art results on the KITTI benchmark.

/pdf/digging-into-self-supervised-monocular-depth-estimation-511yvxepqn.pdf

Digging Into Self-Supervised Monocular Depth Estimation

Per-pixel ground-truth depth data is challenging to acquire at scale. To overcome this limitation, self-supervised learning has emerged as a promising alternative for training models to perform monocular depth estimation. In this paper, we propose a set of improvements, which together result in both quantitatively and qualitatively improved depth maps compared to competing self-supervised methods. 
Research on self-supervised monocular training usually explores increasingly complex architectures, loss functions, and image formation models, all of which have recently helped to close the gap with fully-supervised methods. We show that a surprisingly simple model, and associated design choices, lead to superior predictions. In particular, we propose (i) a minimum reprojection loss, designed to robustly handle occlusions, (ii) a full-resolution multi-scale sampling method that reduces visual artifacts, and (iii) an auto-masking loss to ignore training pixels that violate camera motion assumptions. We demonstrate the effectiveness of each component in isolation, and show high quality, state-of-the-art results on the KITTI benchmark.

In this paper we formulate structure from motion as a learning problem. We train a convolutional network end-to-end to compute depth and camera motion from successive, unconstrained image pairs. The architecture is composed of multiple stacked encoder-decoder networks, the core part being an iterative network that is able to improve its own predictions. The network estimates not only depth and motion, but additionally surface normals, optical flow between the images and confidence of the matching. A crucial component of the approach is a training loss based on spatial relative differences. Compared to traditional two-frame structure from motion methods, results are more accurate and more robust. In contrast to the popular depth-from-single-image networks, DeMoN learns the concept of matching and, thus, better generalizes to structures not seen during training.

/pdf/demon-depth-and-motion-network-for-learning-monocular-stereo-21dth5105y.pdf

DeMoN: Depth and Motion Network for Learning Monocular Stereo

/pdf/demon-depth-and-motion-network-for-learning-monocular-stereo-exsz1z9x1o.pdf

We present a system for keyframe-based dense camera tracking and depth map estimation that is entirely learned. For tracking, we estimate small pose increments between the current camera image and a synthetic viewpoint. This significantly simplifies the learning problem and alleviates the dataset bias for camera motions. Further, we show that generating a large number of pose hypotheses leads to more accurate predictions. For mapping, we accumulate information in a cost volume centered at the current depth estimate. The mapping network then combines the cost volume and the keyframe image to update the depth prediction, thereby effectively making use of depth measurements and image-based priors. Our approach yields state-of-the-art results with few images and is robust with respect to noisy camera poses. We demonstrate that the performance of our 6 DOF tracking competes with RGB-D tracking algorithms.We compare favorably against strong classic and deep learning powered dense depth algorithms.

/pdf/deeptam-deep-tracking-and-mapping-405llkl7as.pdf

DeepTAM: Deep Tracking and Mapping

We present an approach to Lagrangian fluid simulation with a new type of convolutional network. Our networks process sets of moving particles, which describe fluids in space and time. Unlike previous approaches, we do not build an explicit graph structure to connect the particles but use spatial convolutions as the main differentiable operation that relates particles to their neighbors. To this end we present a simple, novel, and effective extension of N-D convolutions to the continuous domain. We show that our network architecture can simulate different materials, generalizes to arbitrary collision geometries, and can be used for inverse problems. In addition, we demonstrate that our continuous convolutions outperform prior formulations in terms of accuracy and speed.

/pdf/lagrangian-fluid-simulation-with-continuous-convolutions-1f07aouv47.pdf

Lagrangian Fluid Simulation with Continuous Convolutions

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

/pdf/cam-convs-camera-aware-multi-scale-convolutions-for-single-4bge7p7zlp.pdf

Benjamin Ummenhofer

Papers

DeMoN: Depth and Motion Network for Learning Monocular Stereo

DeMoN: Depth and Motion Network for Learning Monocular Stereo

DeepTAM: Deep Tracking and Mapping

Lagrangian Fluid Simulation with Continuous Convolutions

CAM-Convs: Camera-Aware Multi-Scale Convolutions for Single-View Depth