We introduce Ego4D, a massive-scale egocentric video dataset and benchmark suite. It offers 3,670 hours of daily-life activity video spanning hundreds of scenarios (household, outdoor, workplace, leisure, etc.) captured by 931 unique camera wearers from 74 worldwide locations and 9 different countries. The approach to collection is designed to uphold rigorous privacy and ethics standards, with consenting participants and robust de-identification procedures where relevant. Ego4D dramatically expands the volume of diverse egocentric video footage publicly available to the research community. Portions of the video are accompanied by audio, 3D meshes of the environment, eye gaze, stereo, and/or synchronized videos from multiple egocentric cameras at the same event. Furthermore, we present a host of new benchmark challenges centered around understanding the first-person visual experience in the past (querying an episodic memory), present (analyzing hand-object manipulation, audio-visual conversation, and social interactions), and future (forecasting activities). By publicly sharing this massive annotated dataset and benchmark suite, we aim to push the frontier of first-person perception.

Camera localization methods based on retrieval, local feature matching, and 3D structure-based pose estimation are accurate but require high storage, are slow, and are not privacy-preserving. A method based on scene landmark detection (SLD) was recently proposed to address these limitations. It involves training a convolutional neural network (CNN) to detect a few predetermined, salient, scene-specific 3D points or landmarks and computing camera pose from the associated 2D–3D correspondences. Although SLD outperformed existing learning-based approaches, it was notably less accurate than 3D structure-based methods. In this paper, we show that the accuracy gap was due to insufficient model capacity and noisy labels during training. To mitigate the capacity issue, we propose to split the landmarks into subgroups and train a separate network for each subgroup. To generate better training labels, we propose using dense reconstructions to estimate visibility of scene landmarks. Finally, we present a compact architecture to improve memory efficiency. Accuracy wise, our approach is on par with state of the art structurebased methods on the INDOOR-6 dataset but runs significantly faster and uses less storage. Code and models can be found at https://github.com/microsoft/SceneLandmarkLocalization.

Modern camera localization methods that use image retrieval, feature matching, and 3D structure-based pose estimation require long-term storage of numerous scene images or a vast amount of image features. This can make them unsuitable for resource constrained VR/AR devices and also raises serious privacy concerns. We present a new learned camera localization technique that eliminates the need to store features or a detailed 3D point cloud. Our key idea is to implicitly encode the appearance of a sparse yet salient set of 3D scene points into a convolutional neural network (CNN) that can detect these scene points in query images whenever they are visible. We refer to these points as scene landmarks. We also show that a CNN can be trained to regress bearing vectors for such landmarks even when they are not within the camera's field-of-view. We demonstrate that the predicted landmarks yield accurate pose estimates and that our method outperforms DSAC*, the state-of-the-art in learned localization. Furthermore, extending HLoc (an accurate method) by combining its correspondences with our predictions boosts its accuracy even further.

In this paper, we study a problem of egocentric scene understanding, i.e., predicting depths and surface normals from an egocentric image. Egocentric scene understanding poses unprecedented challenges: (1) due to large head movements, the images are taken from non-canonical viewpoints (i.e., tilted images) where existing models of geometry prediction do not apply; (2) dynamic foreground objects including hands constitute a large proportion of visual scenes. These challenges limit the performance of the existing models learned from large indoor datasets, such as ScanNet [6] and NYUv2 [36], which comprise predominantly upright images of static scenes. We present a multimodal spatial rectifier that stabilizes the egocentric images to a set of reference directions, which allows learning a coherent visual representation. Unlike unimodal spatial rectifier that often produces excessive perspective warp for egocentric images, the multimodal spatial rectifier learns from multiple directions that can minimize the impact of the perspective warp. To learn visual representations of the dynamic foreground objects, we present a new dataset called EDINA (Egocentric Depth on everyday INdoor Activities) that comprises more than 500K synchronized RGBD frames and gravity directions. Equipped with the multimodal spatial rectifier and the EDINA dataset, our proposed method on single-view depth and surface normal estimation significantly outperforms the baselines not only on our ED-INA dataset, but also on other popular egocentric datasets, such as First Person Hand Action (FPHA) [18] and EPIC-KITCHENS [7].

We introduce Ego4D, a massive-scale egocentric video dataset and benchmark suite. It offers 3,670 hours of dailylife activity video spanning hundreds of scenarios (household, outdoor, workplace, leisure, etc.) captured by 931 unique camera wearers from 74 worldwide locations and 9 different countries. The approach to collection is designed to uphold rigorous privacy and ethics standards, with consenting participants and robust de-identification procedures where relevant. Ego4D dramatically expands the volume of diverse egocentric video footage publicly available to the research community. Portions of the video are accompanied by audio, 3D meshes of the environment, eye gaze, stereo, and/or synchronized videos from multiple egocentric cameras at the same event. Furthermore, we present a host of new benchmark challenges centered around understanding the first-person visual experience in the past (querying an episodic memory), present (analyzing hand-object manipulation, audio-visual conversation, and social interactions), and future (forecasting activities). By publicly sharing this massive annotated dataset and benchmark suite, we aim to push the frontier of first-person perception. Project page: https://ego4d-data.org/

In this paper, we address the problem of estimating dense depth from a sequence of images using deep neural networks. Specifically, we employ a dense-optical-flow network to compute correspondences and then triangulate the point cloud to obtain an initial depth map. Parts of the point cloud, however, may be less accurate than others due to lack of common observations or small parallax. To further increase the triangulation accuracy, we introduce a depth-refinement network (DRN) that optimizes the initial depth map based on the image’s contextual cues. In particular, the DRN contains an iterative refinement module (IRM) that improves the depth accuracy over iterations by refining the deep features. Lastly, the DRN also predicts the uncertainty in the refined depths, which is desirable in applications such as measurement selection for scene reconstruction. We show experimentally that our algorithm outperforms state-of-the-art approaches in terms of depth accuracy, and verify that our predicted uncertainty is highly correlated to the actual depth error.

This paper addresses the problem of learning to complete a scene's depth from sparse depth points and images of indoor scenes. Specifically, we study the case in which the sparse depth is computed from a visual-inertial simultaneous localization and mapping (VI-SLAM) system. The resulting point cloud has low density, it is noisy, and has nonuniform spatial distribution, as compared to the input from active depth sensors, e.g., LiDAR or Kinect. Since the VI-SLAM produces point clouds only over textured areas, we compensate for the missing depth of the low-texture surfaces by leveraging their planar structures and their surface normals which is an important intermediate representation. The pre-trained surface normal network, however, suffers from large performance degradation when there is a significant difference in the viewing direction (especially the roll angle) of the test image as compared to the trained ones. To address this limitation, we use the available gravity estimate from the VI-SLAM to warp the input image to the orientation prevailing in the training dataset. This results in a significant performance gain for the surface normal estimate, and thus the dense depth estimates. Finally, we show that our method outperforms other state-of-the-art approaches both on training (ScanNet [1] and NYUv2 [2]) and testing (collected with Azure Kinect [3]) datasets.

While numerous design methods used in industry help designers rapidly brainstorm design ideas, few help them to use theory in the design process. Behavior change theories can support such design activities as understanding, ideating, sketching, and prototyping. We present the Behavior Change Design Sprint (BCDS), a design process for applying behavior change theories to the design process and for prototyping behavior change technologies. BCDS facilitates the application of theories into the design process through a series of exercises that help designers identify intervention placement and project behavioral outcomes, conduct more focused ideation, and advocate for their design rationale. We present our process to create the sprint and findings from a series of sprint deployments.