An official website of the United States government
Conventional image-motion based methods for structure from motion first compute optical flow, then solve for the 3D motion parameters based on the epipolar constraint, and finally recover the 3D geometry of the scene. However, errors in optical flow due to regularization can lead to large errors in 3D motion and structure. This paper investigates whether performance and consistency can be improved by avoiding optical flow estimation in the early stages of the structure-from-motion pipeline, and it proposes a new direct method based on image gradients (normal flow) only. Our main idea lies in a reformulation of the positive-depth constraint – the basis for estimating egomotion from normal flow – as a continuous piecewise differentiable function, which allows the use of well-known minimization techniques to solve for 3D motion. The 3D motion estimate is then refined and structure estimated adding a regularization based on depth. Experimental comparisons on standard synthetic datasets and the real-world driving benchmark dataset Kitti using three different optic flow algorithms show that the method achieves better accuracy in all but one case. Furthermore, it outperforms existing normal flow based 3D motion estimation techniques. Finally, the recovered 3D geometry is shown to be also very accurate.
more »
« less
DiffPoseNet: Direct Differentiable Camera Pose Estimation
Current deep neural network approaches for camera pose estimation rely on scene structure for 3D motion estimation, but this decreases the robustness and thereby makes cross-dataset generalization difficult. In contrast, classical approaches to structure from motion estimate 3D motion utilizing optical flow and then compute depth. Their accuracy, however, depends strongly on the quality of the optical flow. To avoid this issue, direct methods have been proposed, which separate 3D motion from depth estimation, but compute 3D motion using only image gradients in the form of normal flow. In this paper, we introduce a network NFlowNet, for normal flow estimation which is used to enforce robust and direct constraints. In particular, normal flow is used to estimate relative camera pose based on the cheirality (depth positivity) constraint. We achieve this by formulating the optimization problem as a differentiable cheirality layer, which allows for end-to-end learning of camera pose. We perform extensive qualitative and quantitative evaluation of the proposed DiffPoseNet’s sensitivity to noise and its generalization across datasets. We compare our approach to existing state-of-the-art methods on KITTI, TartanAir, and TUM-RGBD datasets.
more »
« less
- Award ID(s):
- 2020624
- PAR ID:
- 10376891
- Date Published:
- Journal Name:
- IEEE Conference on Computer Vision and Pattern Recognition
- ISSN:
- 2163-6648
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
FlowCam: Training generalizable 3D radiance fields without camera poses via pixel-aligned scene flowReconstruction of 3D neural fields from posed images has emerged as a promising method for self-supervised representation learning. The key challenge preventing the deployment of these 3D scene learners on large-scale video data is their dependence on precise camera poses from structure-from-motion, which is prohibitively expensive to run at scale. We propose a method that jointly reconstructs camera poses and 3D neural scene representations online and in a single forward pass. We estimate poses by first lifting frame-to-frame optical flow to 3D scene flow via differentiable rendering, preserving locality and shift-equivariance of the image processing backbone. SE(3) camera pose estimation is then performed via a weighted least-squares fit to the scene flow field. This formulation enables us to jointly supervise pose estimation and a generalizable neural scene representation via re-rendering the input video, and thus, train end-to-end and fully self-supervised on real-world video datasets. We demonstrate that our method performs robustly on diverse, real-world video, notably on sequences traditionally challenging to optimization-based pose estimation techniques.more » « less
-
FlowCam: Training generalizable 3D radiance fields without camera poses via pixel-aligned scene flowReconstruction of 3D neural fields from posed images has emerged as a promising method for self-supervised representation learning. The key challenge preventing the deployment of these 3D scene learners on large-scale video data is their dependence on precise camera poses from structure-from-motion, which is prohibitively expensive to run at scale. We propose a method that jointly reconstructs camera poses and 3D neural scene representations online and in a single forward pass. We estimate poses by first lifting frame-to-frame optical flow to 3D scene flow via differentiable rendering, preserving locality and shift-equivariance of the image processing backbone. SE(3) camera pose estimation is then performed via a weighted least-squares fit to the scene flow field. This formulation enables us to jointly supervise pose estimation and a generalizable neural scene representation via re-rendering the input video, and thus, train end-to-end and fully self-supervised on real-world video datasets. We demonstrate that our method performs robustly on diverse, real-world video, notably on sequences traditionally challenging to optimization-based pose estimation techniques.more » « less
-
One of the grand challenges in computer vision is to recover 3D poses and shapes of multiple human bodies with absolute scales from a single RGB image. The challenge stems from the inherent depth and scale ambiguity from a single view. The state of the art on 3D human pose and shape estimation mainly focuses on estimating the 3D joint locations relative to the root joint, defined as the pelvis joint. In this paper, a novel approach called Absolute-ROMP is proposed, which builds upon a one-stage multi-person 3D mesh predictor network, ROMP, to estimate multi-person 3D poses and shapes, but with absolute scales from a single RGB image. To achieve this, we introduce absolute root joint localization in the camera coordinate frame, which enables the estimation of 3D mesh coordinates of all persons in the image and their root joint locations normalized by the focal point. Moreover, a CNN and transformer hybrid network, called TransFocal, is proposed to predict the focal length of the image’s camera. This enables Absolute-ROMP to obtain absolute depth information of all joints in the camera coordinate frame, further improving the accuracy of our proposed method. The Absolute-ROMP is evaluated on the root joint localization and root-relative 3D pose estimation tasks on publicly available multi-person 3D pose datasets, and TransFocal is evaluated on a dataset created from the Pano360 dataset. Our proposed approach achieves state-of-the-art results on these tasks, outperforming existing methods or has competitive performance. Due to its real-time performance, our method is applicable to in-the-wild images and videos.more » « less
-
Recovering multi-person 3D poses and shapes with absolute scales from a single RGB image is a challenging task due to the inherent depth and scale ambiguity from a single view. Current works on 3D pose and shape estimation tend to mainly focus on the estimation of the 3D joint locations relative to the root joint , usually defined as the one closest to the shape centroid, in case of humans defined as the pelvis joint. In this paper, we build upon an existing multi-person 3D mesh predictor network, ROMP, to create Absolute-ROMP. By adding absolute root joint localization in the camera coordinate frame, we are able to estimate multi-person 3D poses and shapes with absolute scales from a single RGB image. Such a single-shot approach allows the system to better learn and reason about the inter-person depth relationship, thus improving multi-person 3D estimation. In addition to this end to end network, we also train a CNN and transformer hybrid network, called TransFocal, to predict the f ocal length of the image’s camera. Absolute-ROMP estimates the 3D mesh coordinates of all persons in the image and their root joint locations normalized by the focal point. We then use TransFocal to obtain focal length and get absolute depth information of all joints in the camera coordinate frame. We evaluate Absolute-ROMP on the root joint localization and root-relative 3D pose estimation tasks on publicly available multi-person 3D pose datasets. We evaluate TransFocal on dataset created from the Pano360 dataset and both are applicable to in-the-wild images and videos, due to real time performance.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Conference Paper:
- The DOI is not currently available.
