We introduce HOT3D, a publicly available dataset for egocentric hand and object tracking in 3D. The dataset offers over 833 minutes (3.7M+ images) of recordings that feature 19 subjects interacting with 33 diverse rigid objects. In addition to simple pickup, observe, and put-down actions, the subjects perform actions typical for a kitchen, office, and living room environment. The recordings include multiple synchronized data streams containing egocentric multi-view RGB/monochrome images, eye gaze signal, scene point clouds, and 3D poses of cameras, hands, and objects. The dataset is recorded with two headsets from Meta: Project Aria, which is a research prototype of AI glasses, and Quest 3, a virtual-reality headset that has shipped millions of units. Ground-truth poses were obtained by a motion-capture system using small optical markers attached to hands and objects. Hand annotations are provided in the UmeTrack and MANO formats, and objects are represented by 3D meshes with PBR materials obtained by an in-house scanner. In our experiments, we demonstrate the effectiveness of multi-view egocentric data for three popular tasks: 3D hand tracking, model-based 6DoF object pose estimation, and 3D lifting of unknown in-hand objects. The evaluated multi-view methods, whose benchmarking is uniquely enabled by HOT3D, significantly outperform their single-view counterparts.

AR/VR devices have started to adopt hand tracking, in lieu of controllers, to support user interaction. However, today’s hand input rely primarily on one gesture: pinch. Moreover, current mappings of hand motion to use cases like VR locomotion and content scrolling involve more complex and larger arm motions than joystick or trackpad usage. STMG increases the gesture space by recognizing additional small thumb-based microgestures from skeletal tracking running on a headset. We take a machine learning approach and achieve a 95.1% recognition accuracy across seven thumb gestures performed on the index finger surface: four directional thumb swipes (left, right, forward, backward), thumb tap, and fingertip pinch start and pinch end. We detail the components to our machine learning pipeline and highlight our design decisions and lessons learned in producing a well generalized model. We then demonstrate how these microgestures simplify and reduce arm motions for hand-based locomotion and scrolling interactions.

The task of inputting text within virtual reality has attracted significant research attention over the last five years. Less well explored is the related task of correcting inputted text when errors are made. This is despite the fact that considerable time and frustration stems from efforts to correct text. In this paper, we bridge this gap in prior research and explore efficient methods for supporting text input correction in virtual reality. We present a characterization of the types and frequencies of errors encountered when inputting text in virtual reality and an analysis of effective editing strategies. We also present the results of a user study evaluating the performance and usability trade-offs for several interaction methods leveraging the unique capabilities of modern head-mounted displays.

Hands are the primary means through which humans interact with the world. Reliable and always-available hand pose inference could yield new and intuitive control schemes for human-computer interactions, particularly in virtual and augmented reality. Computer vision is effective but requires one or multiple cameras and can struggle with occlusions, limited field of view, and poor lighting. Wearable wrist-based surface electromyography (sEMG) presents a promising alternative as an always-available modality sensing muscle activities that drive hand motion. However, sEMG signals are strongly dependent on user anatomy and sensor placement; existing sEMG models have thus required hundreds of users and device placements to effectively generalize for tasks other than pose inference. To facilitate progress on sEMG pose inference, we introduce the emg2pose benchmark, which is to our knowledge the first publicly available dataset of high-quality hand pose labels and wrist sEMG recordings. emg2pose contains 2kHz, 16 channel sEMG and pose labels from a 26-camera motion capture rig for 193 users, 370 hours, and 29 stages with diverse gestures - a scale comparable to vision-based hand pose datasets. We provide competitive baselines and challenging tasks evaluating real-world generalization scenarios: held-out users, sensor placements, and stages. This benchmark provides the machine learning community a platform for exploring complex generalization problems, holding potential to significantly enhance the development of sEMG-based human-computer interactions.

Text input is a critical component of any general purpose computing system, yet efficient and natural text input remains a challenge in AR and VR. Headset based hand-tracking has recently become pervasive among consumer VR devices and affords the opportunity to enable touch typing on virtual keyboards. We present an approach for decoding touch typing on uninstrumented flat surfaces using only egocentric camera-based hand-tracking as input. While egocentric hand-tracking accuracy is limited by issues like self occlusion and image fidelity, we show that a sufficiently diverse training set of hand motions paired with typed text can enable a deep learning model to extract signal from this noisy input. Furthermore, by carefully designing a closed-loop data collection process, we can train an end-to-end text decoder that accounts for natural sloppy typing on virtual keyboards. We evaluate our work with a user study (n=18) showing a mean online throughput of 42.4 WPM with an uncorrected error rate (UER) of 7% with our method compared to a physical keyboard baseline of 74.5 WPM at 0.8% UER, showing progress towards unlocking productivity and high throughput use cases in AR/VR.

While passive surfaces offer numerous benefits for interaction in mixed reality, reliably detecting touch input solely from head-mounted cameras has been a long-standing challenge. Camera specifics, hand self-occlusion, and rapid movements of both head and fingers introduce considerable uncertainty about the exact location of touch events. Existing methods have thus not been capable of achieving the performance needed for robust interaction.

Convolutional neural network inference on video input is computationally expensive and requires high memory bandwidth. Recently, DeltaCNN [26] managed to reduce the cost by only processing pixels with significant updates over the previous frame. However, DeltaCNN relies on static camera input. Moving cameras add new challenges in how to fuse newly unveiled image regions with already processed regions efficiently to minimize the update rate - without increasing memory overhead and without knowing the camera extrinsics of future frames. In this work, we propose MotionDeltaCNN, a sparse CNN inference framework that supports moving cameras. We introduce spherical buffers and padded convolutions to enable seamless fusion of newly unveiled regions and previously processed regions – without increasing memory footprint. Our evaluation shows that we outperform DeltaCNN by up to 90% for moving camera videos.

Integrated hand-tracking on modern virtual reality (VR) headsets can be readily exploited to deliver mid-air virtual input surfaces for text entry. These virtual input surfaces can closely replicate the experience of typing on a Qwerty keyboard on a physical touchscreen, thereby allowing users to leverage their pre-existing typing skills. However, the lack of passive haptic feedback, unconstrained user motion, and potential tracking inaccuracies or observability issues encountered in this interaction setting typically degrades the accuracy of user articulations. We present a comprehensive exploration of error-tolerant probabilistic hand-based input methods to support effective text input on a mid-air virtual Qwerty keyboard. Over three user studies we examine the performance potential of hand-based text input under both gesture and touch typing paradigms. We demonstrate typical entry rates in the range of 20 to 30 wpm and average peak entry rates of 40 to 45 wpm.

Convolutional neural network inference on video data requires powerful hardware for real-time processing. Given the inherent coherence across consecutive frames, large parts of a video typically change little. By skipping identical image regions and truncating insignificant pixel updates, computational redundancy can in theory be reduced significantly. However, these theoretical savings have been difficult to translate into practice, as sparse updates hamper computational consistency and memory access coherence; which are key for efficiency on real hardware. With DeltaCNN, we present a sparse convolutional neural network framework that enables sparse frame-by-frame updates to accelerate video inference in practice. We provide sparse implementations for all typical CNN layers and propagate sparse feature updates end-to-end – without accumulating errors over time. DeltaCNN is applicable to all convolutional neural networks without retraining. To the best of our knowledge, we are the first to significantly outperform the dense reference, cuDNN, in practical settings, achieving speedups of up to 7x with only marginal differences in accuracy. Our CUDA kernels and PyTorch extensions can be found at https://github.com/facebookresearch/DeltaCNN.

Assembly101 is a new procedural activity dataset fea-turing 4321 videos of people assembling and disassembling 101 “take-apart” toy vehicles. Participants work without fixed instructions, and the sequences feature rich and natu-ral variations in action ordering, mistakes, and corrections. Assembly101 is the first multi-view action dataset, with si-multaneous static (8) and egocentric (4) recordings. Se-quences are annotated with more than 100K coarse and 1M fine-grained action segments, and I8M 3D hand poses. We benchmark on three action understanding tasks: recognition, anticipation and temporal segmentation. Ad-ditionally, we propose a novel task of detecting mistakes. The unique recording format and rich set of annotations al-low us to investigate generalization to new toys, cross-view transfer, long-tailed distributions, and pose vs. appearance. We envision that Assemblyl0l will serve as a new challenge to investigate various activity understanding problems.

Real-time tracking of 3D hand pose in world space is a challenging problem and plays an important role in VR interaction. Existing work in this space are limited to either producing root-relative (versus world space) 3D pose or rely on multiple stages such as generating heatmaps and kinematic optimization to obtain 3D pose. Moreover, the typical VR scenario, which involves multi-view tracking from wide field of view (FOV) cameras is seldom addressed by these methods. In this paper, we present a unified end-to-end differentiable framework for multi-view, multi-frame hand tracking that directly predicts 3D hand pose in world space. We demonstrate the benefits of end-to-end differentiabilty by extending our framework with downstream tasks such as jitter reduction and pinch prediction. To demonstrate the efficacy of our model, we further present a new large-scale egocentric hand pose dataset that consists of both real and synthetic data. Experiments show that our system trained on this dataset handles various challenging interactive motions, and has been successfully applied to real-time VR applications.

Tracking body and hand motions in 3D space is essential for social and self-presence in augmented and virtual environments. Unlike the popular 3D pose estimation setting, the problem is often formulated as egocentric tracking based on embodied perception (e.g., egocentric cameras, handheld sensors). In this article, we propose a new data-driven framework for egocentric body tracking, targeting challenges of omnipresent occlusions in optimization-based methods (e.g., inverse kinematics solvers). We first collect a large-scale motion capture dataset with both body and finger motions using optical markers and inertial sensors. This dataset focuses on social scenarios and captures ground truth poses under self-occlusions and body-hand interactions. We then simulate the occlusion patterns in head-mounted camera views on the captured ground truth using a ray casting algorithm and learn a deep neural network to infer the occluded body parts. Our experiments show that our method is able to generate high-fidelity embodied poses by applying the proposed method to the task of real-time egocentric body tracking, finger motion synthesis, and 3-point inverse kinematics.

Fully immersive experiences in AR/VR depend on re-constructing the full body pose of the user without restricting their motion. In this paper we study the use of body-worn electromagnetic (EM) field-based sensing for the task of 3D human pose reconstruction. To this end, we present a method to estimate SMPL parameters from 6-12 EM sensors. We leverage a customized wearable system consisting of wireless EM sensors measuring time-synchronized 6D poses at 120 Hz. To provide accurate poses even with little user instrumentation, we adopt a recently proposed hybrid framework, learned gradient descent (LGD), to iteratively estimate SMPL pose and shape from our input measurements. This allows us to harness powerful pose priors to cope with the idiosyncrasies of the input data and achieve accurate pose estimates. The proposed method uses AMASS to synthesize virtual EM-sensor data and we show that it generalizes well to a newly captured real dataset consisting of a total of 36 minutes of motion from 5 subjects. We achieve reconstruction errors as low as 31.8 mm and 13.3 degrees, outperforming both pure learning- and pure optimization-based methods. Code and data is available under https://ait.ethz.ch/projects/2021/em-pose.

We present a lightweight solution to recover 3D pose from multi-view images captured with spatially calibrated cameras. Building upon recent advances in interpretable representation learning, we exploit 3D geometry to fuse input images into a unified latent representation of pose, which is disentangled from camera view-points. This allows us to reason effectively about 3D pose across different views without using compute-intensive volumetric grids. Our architecture then conditions the learned representation on camera projection operators to produce accurate per-view 2d detections, that can be simply lifted to 3D via a differentiable Direct Linear Transform (DLT) layer. In order to do it efficiently, we propose a novel implementation of DLT that is orders of magnitude faster on GPU architectures than standard SVD-based triangulation methods. We evaluate our approach on two large-scale human pose datasets (H36M and Total Capture): our method outperforms or performs comparably to the state-of-the-art volumetric methods, while, unlike them, yielding real-time performance.

We propose a novel text decoding method that enables touch typing on an uninstrumented flat surface. Rather than relying on physical keyboards or capacitive touch, our method takes as input hand motion of the typist, obtained through hand-tracking, and decodes this motion directly into text. We use a temporal convolutional network to represent a motion model that maps the hand motion, represented as a sequence of hand pose features, into text characters. To enable touch typing without the haptic feedback of a physical keyboard, we had to address more erratic typing motion due to drift of the fingers. Thus, we incorporate a language model as a text prior and use beam search to efficiently combine our motion and language models to decode text from erratic or ambiguous hand motion. We collected a dataset of 20 touch typists and evaluated our model on several baselines, including contact-based text decoding and typing on a physical keyboard. Our proposed method is able to leverage continuous hand pose information to decode text more accurately than contact-based methods and an offline study shows parity (73 WPM, 2.38% UER) with typing on a physical keyboard. Our results show that hand-tracking has the potential to enable rapid text entry in mobile environments.

We present a system for real-time hand-tracking to drive virtual and augmented reality (VR/AR) experiences. Using four fisheye monochrome cameras, our system generates accurate and low-jitter 3D hand motion across a large working volume for a diverse set of users. We achieve this by proposing neural network architectures for detecting hands and estimating hand keypoint locations. Our hand detection network robustly handles a variety of real world environments. The keypoint estimation network leverages tracking history to produce spatially and temporally consistent poses. We design scalable, semi-automated mechanisms to collect a large and diverse set of ground truth data using a combination of manual annotation and automated tracking. Additionally, we introduce a detection-by-tracking method that increases smoothness while reducing the computational cost; the optimized system runs at 60Hz on PC and 30Hz on a mobile processor. Together, these contributions yield a practical system for capturing a user's hands and is the default feature on the Oculus Quest VR headset powering input and social presence.

Optical marker-based motion capture is the dominant way for obtaining high-fidelity human body animation for special effects, movies, and video games. However, motion capture has seen limited application to the human hand due to the difficulty of automatically identifying (or labeling) identical markers on self-similar fingers. We propose a technique that frames the labeling problem as a keypoint regression problem conducive to a solution using convolutional neural networks. We demonstrate robustness of our labeling solution to occlusion, ghost markers, hand shape, and even motions involving two hands or handheld objects. Our technique is equally applicable to sparse or dense marker sets and can run in real-time to support interaction prototyping with high-fidelity hand tracking and hand presence in virtual reality.

We propose a system for real-time six degrees of freedom (6DoF) tracking of a passive stylus that achieves sub-millimeter accuracy, which is suitable for writing or drawing in mixed reality applications. Our system is particularly easy to implement, requiring only a monocular camera, a 3D printed dodecahedron, and hand-glued binary square markers. The accuracy and performance we achieve are due to model-based tracking using a calibrated model and a combination of sparse pose estimation and dense alignment. We demonstrate the system performance in terms of speed and accuracy on a number of synthetic and real datasets, showing that it can be competitive with state-of-the-art multi-camera motion capture systems. We also demonstrate several applications of the technology ranging from 2D and 3D drawing in VR to general object manipulation and board games.