Lingni Ma

Jakob Engel, Kiran Somasundaram, Michael Goesele et al. · mit

Egocentric, multi-modal data as available on future augmented reality (AR) devices provides unique challenges and opportunities for machine perception. These future devices will need to be all-day wearable in a socially acceptable form-factor to support always available, context-aware and personalized AI applications. Our team at Meta Reality Labs Research built the Aria device, an egocentric, multi-modal data recording and streaming device with the goal to foster and accelerate research in this area. In this paper, we describe the Aria device hardware including its sensor configuration and the corresponding software tools that enable recording and processing of such data.

Shihao Zou, Yuanlu Xu, Chao Li et al. · meta-ai

Multi-person pose understanding from RGB videos involves three complex tasks: pose estimation, tracking and motion forecasting. Intuitively, accurate multi-person pose estimation facilitates robust tracking, and robust tracking builds crucial history for correct motion forecasting. Most existing works either focus on a single task or employ multi-stage approaches to solving multiple tasks separately, which tends to make sub-optimal decision at each stage and also fail to exploit correlations among the three tasks. In this paper, we propose Snipper, a unified framework to perform multi-person 3D pose estimation, tracking, and motion forecasting simultaneously in a single stage. We propose an efficient yet powerful deformable attention mechanism to aggregate spatiotemporal information from the video snippet. Building upon this deformable attention, a video transformer is learned to encode the spatiotemporal features from the multi-frame snippet and to decode informative pose features for multi-person pose queries. Finally, these pose queries are regressed to predict multi-person pose trajectories and future motions in a single shot. In the experiments, we show the effectiveness of Snipper on three challenging public datasets where our generic model rivals specialized state-of-art baselines for pose estimation, tracking, and forecasting.

Fangyin Wei, Rohan Chabra, Lingni Ma et al.

Learning geometry, motion, and appearance priors of object classes is important for the solution of a large variety of computer vision problems. While the majority of approaches has focused on static objects, dynamic objects, especially with controllable articulation, are less explored. We propose a novel approach for learning a representation of the geometry, appearance, and motion of a class of articulated objects given only a set of color images as input. In a self-supervised manner, our novel representation learns shape, appearance, and articulation codes that enable independent control of these semantic dimensions. Our model is trained end-to-end without requiring any articulation annotations. Experiments show that our approach performs well for different joint types, such as revolute and prismatic joints, as well as different combinations of these joints. Compared to state of the art that uses direct 3D supervision and does not output appearance, we recover more faithful geometry and appearance from 2D observations only. In addition, our representation enables a large variety of applications, such as few-shot reconstruction, the generation of novel articulations, and novel view-synthesis.

Shreyas Hampali, Tomas Hodan, Luan Tran et al.

We propose a method for in-hand 3D scanning of an unknown object with a monocular camera. Our method relies on a neural implicit surface representation that captures both the geometry and the appearance of the object, however, by contrast with most NeRF-based methods, we do not assume that the camera-object relative poses are known. Instead, we simultaneously optimize both the object shape and the pose trajectory. As direct optimization over all shape and pose parameters is prone to fail without coarse-level initialization, we propose an incremental approach that starts by splitting the sequence into carefully selected overlapping segments within which the optimization is likely to succeed. We reconstruct the object shape and track its poses independently within each segment, then merge all the segments before performing a global optimization. We show that our method is able to reconstruct the shape and color of both textured and challenging texture-less objects, outperforms classical methods that rely only on appearance features, and that its performance is close to recent methods that assume known camera poses.

Enric Corona, Tomas Hodan, Minh Vo et al.

This paper proposes a do-it-all neural model of human hands, named LISA. The model can capture accurate hand shape and appearance, generalize to arbitrary hand subjects, provide dense surface correspondences, be reconstructed from images in the wild and easily animated. We train LISA by minimizing the shape and appearance losses on a large set of multi-view RGB image sequences annotated with coarse 3D poses of the hand skeleton. For a 3D point in the hand local coordinate, our model predicts the color and the signed distance with respect to each hand bone independently, and then combines the per-bone predictions using predicted skinning weights. The shape, color and pose representations are disentangled by design, allowing to estimate or animate only selected parameters. We experimentally demonstrate that LISA can accurately reconstruct a dynamic hand from monocular or multi-view sequences, achieving a noticeably higher quality of reconstructed hand shapes compared to baseline approaches. Project page: https://www.iri.upc.edu/people/ecorona/lisa/.

Lin Huang, Tomas Hodan, Lingni Ma et al.

We propose a method for estimating the 6DoF pose of a rigid object with an available 3D model from a single RGB image. Unlike classical correspondence-based methods which predict 3D object coordinates at pixels of the input image, the proposed method predicts 3D object coordinates at 3D query points sampled in the camera frustum. The move from pixels to 3D points, which is inspired by recent PIFu-style methods for 3D reconstruction, enables reasoning about the whole object, including its (self-)occluded parts. For a 3D query point associated with a pixel-aligned image feature, we train a fully-connected neural network to predict: (i) the corresponding 3D object coordinates, and (ii) the signed distance to the object surface, with the first defined only for query points in the surface vicinity. We call the mapping realized by this network as Neural Correspondence Field. The object pose is then robustly estimated from the predicted 3D-3D correspondences by the Kabsch-RANSAC algorithm. The proposed method achieves state-of-the-art results on three BOP datasets and is shown superior especially in challenging cases with occlusion. The project website is at: linhuang17.github.io/NCF.

Evin Pınar Örnek, Yann Labbé, Bugra Tekin et al.

We propose FoundPose, a model-based method for 6D pose estimation of unseen objects from a single RGB image. The method can quickly onboard new objects using their 3D models without requiring any object- or task-specific training. In contrast, existing methods typically pre-train on large-scale, task-specific datasets in order to generalize to new objects and to bridge the image-to-model domain gap. We demonstrate that such generalization capabilities can be observed in a recent vision foundation model trained in a self-supervised manner. Specifically, our method estimates the object pose from image-to-model 2D-3D correspondences, which are established by matching patch descriptors from the recent DINOv2 model between the image and pre-rendered object templates. We find that reliable correspondences can be established by kNN matching of patch descriptors from an intermediate DINOv2 layer. Such descriptors carry stronger positional information than descriptors from the last layer, and we show their importance when semantic information is ambiguous due to object symmetries or a lack of texture. To avoid establishing correspondences against all object templates, we develop an efficient template retrieval approach that integrates the patch descriptors into the bag-of-words representation and can promptly propose a handful of similarly looking templates. Additionally, we apply featuremetric alignment to compensate for discrepancies in the 2D-3D correspondences caused by coarse patch sampling. The resulting method noticeably outperforms existing RGB methods for refinement-free pose estimation on the standard BOP benchmark with seven diverse datasets and can be seamlessly combined with an existing render-and-compare refinement method to achieve RGB-only state-of-the-art results. Project page: evinpinar.github.io/foundpose.

Qiao Gu, Lingni Ma, Adam W Harley et al.

Controllable and physically grounded egocentric video generation is essential for embodied agents to reason about how their own and others' actions manifest and change the world. Compared to generic video synthesis, egocentric generation is especially challenging: the camera is tightly coupled to the actor, leading to rapid viewpoint changes and frequent self-occlusions; the underlying actions are subtle, articulated, and often only partially visible; and both the people and the scene state must evolve consistently with the specified controls. We present E$^3$C, a controllable video diffusion framework for egocentric generation that builds structured and compact conditions disentangling persistent scene structure from human-driven dynamics. From context frames, E$^3$C constructs a semi-dense point cloud-based 3D memory and augments each point with appearance descriptors from video-VAE features. Rendering this memory into target viewpoints produces conditioning aligned with the target frames. Human dynamics are modeled separately. The observed people in the scene are controlled by skeleton renderings (exo human control), while the camera wearer is specified by their 3D body joints and 6DoF wrist motion (ego human control). To preserve ego human control when the wearer's body parts are invisible, we introduce an ego motion encoder that produces persistent cross-attention tokens. Experiments on Nymeria show that E$^3$C improves visual fidelity, camera-motion accuracy, object consistency, and ego & exo human control over strong baselines, while also enabling intuitive scene editing.

Vladimir Guzov, Yifeng Jiang, Fangzhou Hong et al.

This paper investigates the generation of realistic full-body human motion using a single head-mounted device with an outward-facing color camera and the ability to perform visual SLAM. To address the ambiguity of this setup, we present HMD^2, a novel system that balances motion reconstruction and generation. From a reconstruction standpoint, it aims to maximally utilize the camera streams to produce both analytical and learned features, including head motion, SLAM point cloud, and image embeddings. On the generative front, HMD^2 employs a multi-modal conditional motion diffusion model with a Transformer backbone to maintain temporal coherence of generated motions, and utilizes autoregressive inpainting to facilitate online motion inference with minimal latency (0.17 seconds). We show that our system provides an effective and robust solution that scales to a diverse dataset of over 200 hours of motion in complex indoor and outdoor environments.

Fangzhou Hong, Vladimir Guzov, Hyo Jin Kim et al.

As the prevalence of wearable devices, learning egocentric motions becomes essential to develop contextual AI. In this work, we present EgoLM, a versatile framework that tracks and understands egocentric motions from multi-modal inputs, e.g., egocentric videos and motion sensors. EgoLM exploits rich contexts for the disambiguation of egomotion tracking and understanding, which are ill-posed under single modality conditions. To facilitate the versatile and multi-modal framework, our key insight is to model the joint distribution of egocentric motions and natural languages using large language models (LLM). Multi-modal sensor inputs are encoded and projected to the joint latent space of language models, and used to prompt motion generation or text generation for egomotion tracking or understanding, respectively. Extensive experiments on large-scale multi-modal human motion dataset validate the effectiveness of EgoLM as a generalist model for universal egocentric learning.

Daniel DeTone, Tianwei Shen, Fan Zhang et al.

Detecting and localizing objects in space is a fundamental computer vision problem. While much progress has been made to solve 2D object detection, 3D object localization is much less explored and far from solved, especially for open-world categories. To address this research challenge, we propose Boxer, an algorithm to estimate static 3D bounding boxes (3DBBs) from 2D open-vocabulary object detections, posed images and optional depth either represented as a sparse point cloud or dense depth. At its core is BoxerNet, a transformer-based network which lifts 2D bounding box (2DBB) proposals into 3D, followed by multi-view fusion and geometric filtering to produce globally consistent de-duplicated 3DBBs in metric world space. Boxer leverages the power of existing 2DBB detection algorithms (e.g. DETIC, OWLv2, SAM3) to localize objects in 2D. This allows the main BoxerNet model to focus on lifting to 3D rather than detecting, ultimately reducing the demand for costly annotated 3DBB training data. Extending the CuTR formulation, we incorporate an aleatoric uncertainty for robust regression, a median depth patch encoding to support sparse depth inputs, and large-scale training with over 1.2 million unique 3DBBs. BoxerNet outperforms state-of-the-art baselines in open-world 3DBB lifting, including CuTR in egocentric settings without dense depth (0.532 vs. 0.010 mAP) and on CA-1M with dense depth available (0.412 vs. 0.250 mAP).

Daniel DeTone, Federica Bogo, Eric-Tuan Le et al.

The Nymeria Dataset, released in 2024, is a large-scale collection of in-the-wild human activities captured with multiple egocentric wearable devices that are spatially localized and temporally synchronized. It provides body-motion ground truth recorded with a motion-capture suit, device trajectories, semi-dense 3D point clouds, and in-context narrations. In this paper, we upgrade Nymeria and introduce NymeriaPlus. NymeriaPlus features: (1) improved human motion in Momentum Human Rig (MHR) and SMPL formats; (2) dense 3D and 2D bounding box annotations for indoor objects and structural elements; (3) instance-level 3D object reconstructions; and (4) additional modalities e.g., basemap recordings, audio, and wristband videos. By consolidating these complementary modalities and annotations into a single, coherent benchmark, NymeriaPlus strengthens Nymeria into a more powerful in-the-wild egocentric dataset. We expect NymeriaPlus to bridge a key gap in existing egocentric resources and to support a broader range of research, including unique explorations of multimodal learning for embodied AI.

Nan Yang, Julian Straub, Fan Zhang et al.

Tracking 3D human motion from egocentric multi-camera headset is challenged by severe egomotion, partial visibility or occlusions and lack of training data. Existing methods designed for monocular video often require static or slowly-moving cameras and cannot efficiently leverage multi-view, calibrated and localized input. This makes them brittle and prone to fail on dynamic egocentric captures. We propose LAMP (Localization Aware Multi-camera People Tracking): a novel, simple framework to solve this via early disentanglement of observer and target motion. LAMP introduces a two-step process. First, we leverage the known device 6 DoF motion and calibration to convert detected 2D body keypoints from all cameras over a temporal window into a unified 3D world reference frame. Second, an end-to-end-trained spatio-temporal transformer fits 3D human motion directly to this 3D ray cloud. This "lift-then-fit" approach allows LAMP to learn and leverage a natural human motion prior in the world-space, as well as providing an elegant framework to flexibly incorporate information from multiple temporally asynchronous, partially observing and moving cameras. LAMP achieves state-of-the-art results on monocular benchmarks, while significantly outperforming baselines for our targeted egocentric setting.

Lingni Ma, Yuting Ye, Fangzhou Hong et al.

We introduce Nymeria - a large-scale, diverse, richly annotated human motion dataset collected in the wild with multiple multimodal egocentric devices. The dataset comes with a) full-body ground-truth motion; b) multiple multimodal egocentric data from Project Aria devices with videos, eye tracking, IMUs and etc; and c) a third-person perspective by an additional observer. All devices are precisely synchronized and localized in on metric 3D world. We derive hierarchical protocol to add in-context language descriptions of human motion, from fine-grain motion narration, to simplified atomic action and high-level activity summarization. To the best of our knowledge, Nymeria dataset is the world's largest collection of human motion in the wild; first of its kind to provide synchronized and localized multi-device multimodal egocentric data; and the world's largest motion-language dataset. It provides 300 hours of daily activities from 264 participants across 50 locations, total travelling distance over 399Km. The language descriptions contain 301.5K sentences in 8.64M words from a vocabulary size of 6545. To demonstrate the potential of the dataset, we evaluate several SOTA algorithms for egocentric body tracking, motion synthesis, and action recognition. Data and code are open-sourced for research (c.f. https://www.projectaria.com/datasets/nymeria).

Dongseok Yang, Jiho Kang, Lingni Ma et al.

Full-body avatar presence is crucial for immersive social and environmental interactions in digital reality. However, current devices only provide three six degrees of freedom (DOF) poses from the headset and two controllers (i.e. three-point trackers). Because it is a highly under-constrained problem, inferring full-body pose from these inputs is challenging, especially when supporting the full range of body proportions and use cases represented by the general population. In this paper, we propose a deep learning framework, DivaTrack, which outperforms existing methods when applied to diverse body sizes and activities. We augment the sparse three-point inputs with linear accelerations from Inertial Measurement Units (IMU) to improve foot contact prediction. We then condition the otherwise ambiguous lower-body pose with the predictions of foot contact and upper-body pose in a two-stage model. We further stabilize the inferred full-body pose in a wide range of configurations by learning to blend predictions that are computed in two reference frames, each of which is designed for different types of motions. We demonstrate the effectiveness of our design on a large dataset that captures 22 subjects performing challenging locomotion for three-point tracking, including lunges, hula-hooping, and sitting. As shown in a live demo using the Meta VR headset and Xsens IMUs, our method runs in real-time while accurately tracking a user's motion when they perform a diverse set of movements.

Rawal Khirodkar, Aayush Bansal, Lingni Ma et al.

We present EgoHumans, a new multi-view multi-human video benchmark to advance the state-of-the-art of egocentric human 3D pose estimation and tracking. Existing egocentric benchmarks either capture single subject or indoor-only scenarios, which limit the generalization of computer vision algorithms for real-world applications. We propose a novel 3D capture setup to construct a comprehensive egocentric multi-human benchmark in the wild with annotations to support diverse tasks such as human detection, tracking, 2D/3D pose estimation, and mesh recovery. We leverage consumer-grade wearable camera-equipped glasses for the egocentric view, which enables us to capture dynamic activities like playing tennis, fencing, volleyball, etc. Furthermore, our multi-view setup generates accurate 3D ground truth even under severe or complete occlusion. The dataset consists of more than 125k egocentric images, spanning diverse scenes with a particular focus on challenging and unchoreographed multi-human activities and fast-moving egocentric views. We rigorously evaluate existing state-of-the-art methods and highlight their limitations in the egocentric scenario, specifically on multi-human tracking. To address such limitations, we propose EgoFormer, a novel approach with a multi-stream transformer architecture and explicit 3D spatial reasoning to estimate and track the human pose. EgoFormer significantly outperforms prior art by 13.6% IDF1 on the EgoHumans dataset.

Binbin Xu, Lingni Ma, Yuting Ye et al.

Neural shape models can represent complex 3D shapes with a compact latent space. When applied to dynamically deforming shapes such as the human hands, however, they would need to preserve temporal coherence of the deformation as well as the intrinsic identity of the subject. These properties are difficult to regularize with manually designed loss functions. In this paper, we learn a neural deformation model that disentangles the identity-induced shape variations from pose-dependent deformations using implicit neural functions. We perform template-free unsupervised learning on 3D scans without explicit mesh correspondence or semantic correspondences of shapes across subjects. We can then apply the learned model to reconstruct partial dynamic 4D scans of novel subjects performing unseen actions. We propose two methods to integrate global pose alignment with our neural deformation model. Experiments demonstrate the efficacy of our method in the disentanglement of identities and pose. Our method also outperforms traditional skeleton-driven models in reconstructing surface details such as palm prints or tendons without limitations from a fixed template.

Miao Liu, Lingni Ma, Kiran Somasundaram et al.

Given a video captured from a first person perspective and the environment context of where the video is recorded, can we recognize what the person is doing and identify where the action occurs in the 3D space? We address this challenging problem of jointly recognizing and localizing actions of a mobile user on a known 3D map from egocentric videos. To this end, we propose a novel deep probabilistic model. Our model takes the inputs of a Hierarchical Volumetric Representation (HVR) of the 3D environment and an egocentric video, infers the 3D action location as a latent variable, and recognizes the action based on the video and contextual cues surrounding its potential locations. To evaluate our model, we conduct extensive experiments on the subset of Ego4D dataset, in which both human naturalistic actions and photo-realistic 3D environment reconstructions are captured. Our method demonstrates strong results on both action recognition and 3D action localization across seen and unseen environments. We believe our work points to an exciting research direction in the intersection of egocentric vision, and 3D scene understanding.

Kejie Li, Martin Rünz, Meng Tang et al.

Object-oriented maps are important for scene understanding since they jointly capture geometry and semantics, allow individual instantiation and meaningful reasoning about objects. We introduce FroDO, a method for accurate 3D reconstruction of object instances from RGB video that infers object location, pose and shape in a coarse-to-fine manner. Key to FroDO is to embed object shapes in a novel learnt space that allows seamless switching between sparse point cloud and dense DeepSDF decoding. Given an input sequence of localized RGB frames, FroDO first aggregates 2D detections to instantiate a category-aware 3D bounding box per object. A shape code is regressed using an encoder network before optimizing shape and pose further under the learnt shape priors using sparse and dense shape representations. The optimization uses multi-view geometric, photometric and silhouette losses. We evaluate on real-world datasets, including Pix3D, Redwood-OS, and ScanNet, for single-view, multi-view, and multi-object reconstruction.

Julian Straub, Thomas Whelan, Lingni Ma et al.

We introduce Replica, a dataset of 18 highly photo-realistic 3D indoor scene reconstructions at room and building scale. Each scene consists of a dense mesh, high-resolution high-dynamic-range (HDR) textures, per-primitive semantic class and instance information, and planar mirror and glass reflectors. The goal of Replica is to enable machine learning (ML) research that relies on visually, geometrically, and semantically realistic generative models of the world - for instance, egocentric computer vision, semantic segmentation in 2D and 3D, geometric inference, and the development of embodied agents (virtual robots) performing navigation, instruction following, and question answering. Due to the high level of realism of the renderings from Replica, there is hope that ML systems trained on Replica may transfer directly to real world image and video data. Together with the data, we are releasing a minimal C++ SDK as a starting point for working with the Replica dataset. In addition, Replica is `Habitat-compatible', i.e. can be natively used with AI Habitat for training and testing embodied agents.

Lingni Ma, Jörg Stückler, Tao Wu et al.

Dense pixelwise prediction such as semantic segmentation is an up-to-date challenge for deep convolutional neural networks (CNNs). Many state-of-the-art approaches either tackle the loss of high-resolution information due to pooling in the encoder stage, or use dilated convolutions or high-resolution lanes to maintain detailed feature maps and predictions. Motivated by the structural analogy between multi-resolution wavelet analysis and the pooling/unpooling layers of CNNs, we introduce discrete wavelet transform (DWT) into the CNN encoder-decoder architecture and propose WCNN. The high-frequency wavelet coefficients are computed at encoder, which are later used at the decoder to unpooled jointly with coarse-resolution feature maps through the inverse DWT. The DWT/iDWT is further used to develop two wavelet pyramids to capture the global context, where the multi-resolution DWT is applied to successively reduce the spatial resolution and increase the receptive field. Experiment with the Cityscape dataset, the proposed WCNNs are computationally efficient and yield improvements the accuracy for high-resolution dense pixelwise prediction.

Lingni Ma, Jörg Stückler, Christian Kerl et al.

Visual scene understanding is an important capability that enables robots to purposefully act in their environment. In this paper, we propose a novel approach to object-class segmentation from multiple RGB-D views using deep learning. We train a deep neural network to predict object-class semantics that is consistent from several view points in a semi-supervised way. At test time, the semantics predictions of our network can be fused more consistently in semantic keyframe maps than predictions of a network trained on individual views. We base our network architecture on a recent single-view deep learning approach to RGB and depth fusion for semantic object-class segmentation and enhance it with multi-scale loss minimization. We obtain the camera trajectory using RGB-D SLAM and warp the predictions of RGB-D images into ground-truth annotated frames in order to enforce multi-view consistency during training. At test time, predictions from multiple views are fused into keyframes. We propose and analyze several methods for enforcing multi-view consistency during training and testing. We evaluate the benefit of multi-view consistency training and demonstrate that pooling of deep features and fusion over multiple views outperforms single-view baselines on the NYUDv2 benchmark for semantic segmentation. Our end-to-end trained network achieves state-of-the-art performance on the NYUDv2 dataset in single-view segmentation as well as multi-view semantic fusion.

Maksym Dzitsiuk, Jürgen Sturm, Robert Maier et al.

Creating 3D maps on robots and other mobile devices has become a reality in recent years. Online 3D reconstruction enables many exciting applications in robotics and AR/VR gaming. However, the reconstructions are noisy and generally incomplete. Moreover, during onine reconstruction, the surface changes with every newly integrated depth image which poses a significant challenge for physics engines and path planning algorithms. This paper presents a novel, fast and robust method for obtaining and using information about planar surfaces, such as walls, floors, and ceilings as a stage in 3D reconstruction based on Signed Distance Fields. Our algorithm recovers clean and accurate surfaces, reduces the movement of individual mesh vertices caused by noise during online reconstruction and fills in the occluded and unobserved regions. We implemented and evaluated two different strategies to generate plane candidates and two strategies for merging them. Our implementation is optimized to run in real-time on mobile devices such as the Tango tablet. In an extensive set of experiments, we validated that our approach works well in a large number of natural environments despite the presence of significant amount of occlusion, clutter and noise, which occur frequently. We further show that plane fitting enables in many cases a meaningful semantic segmentation of real-world scenes.