Xinshuo Weng

Jinkun Cao, Jiangmiao Pang, Xinshuo Weng et al. · cmu

Kalman filter (KF) based methods for multi-object tracking (MOT) make an assumption that objects move linearly. While this assumption is acceptable for very short periods of occlusion, linear estimates of motion for prolonged time can be highly inaccurate. Moreover, when there is no measurement available to update Kalman filter parameters, the standard convention is to trust the priori state estimations for posteriori update. This leads to the accumulation of errors during a period of occlusion. The error causes significant motion direction variance in practice. In this work, we show that a basic Kalman filter can still obtain state-of-the-art tracking performance if proper care is taken to fix the noise accumulated during occlusion. Instead of relying only on the linear state estimate (i.e., estimation-centric approach), we use object observations (i.e., the measurements by object detector) to compute a virtual trajectory over the occlusion period to fix the error accumulation of filter parameters during the occlusion period. This allows more time steps to correct errors accumulated during occlusion. We name our method Observation-Centric SORT (OC-SORT). It remains Simple, Online, and Real-Time but improves robustness during occlusion and non-linear motion. Given off-the-shelf detections as input, OC-SORT runs at 700+ FPS on a single CPU. It achieves state-of-the-art on multiple datasets, including MOT17, MOT20, KITTI, head tracking, and especially DanceTrack where the object motion is highly non-linear. The code and models are available at \url{https://github.com/noahcao/OC_SORT}.

Cheng-hsin Wuu, Ningyuan Zheng, Scott Ardisson et al. · cmu

Photorealistic avatars of human faces have come a long way in recent years, yet research along this area is limited by a lack of publicly available, high-quality datasets covering both, dense multi-view camera captures, and rich facial expressions of the captured subjects. In this work, we present Multiface, a new multi-view, high-resolution human face dataset collected from 13 identities at Reality Labs Research for neural face rendering. We introduce Mugsy, a large scale multi-camera apparatus to capture high-resolution synchronized videos of a facial performance. The goal of Multiface is to close the gap in accessibility to high quality data in the academic community and to enable research in VR telepresence. Along with the release of the dataset, we conduct ablation studies on the influence of different model architectures toward the model's interpolation capacity of novel viewpoint and expressions. With a conditional VAE model serving as our baseline, we found that adding spatial bias, texture warp field, and residual connections improves performance on novel view synthesis. Our code and data is available at: https://github.com/facebookresearch/multiface

Yan Wang, Wenjie Luo, Junjie Bai et al. · nvidia

End-to-end architectures trained via imitation learning have advanced autonomous driving by scaling model size and data, yet performance remains brittle in safety-critical long-tail scenarios where supervision is sparse and causal understanding is limited. To address this, we introduce Alpamayo-R1 (AR1), a vision-language-action model (VLA) that integrates Chain of Causation reasoning with trajectory planning to enhance decision-making in complex driving scenarios. Our approach features three key innovations: (1) the Chain of Causation (CoC) dataset, built through a hybrid auto-labeling and human-in-the-loop pipeline producing decision-grounded, causally linked reasoning traces aligned with driving behaviors; (2) a modular VLA architecture combining Cosmos-Reason, a Vision-Language Model pre-trained for Physical AI applications, with a diffusion-based trajectory decoder that generates dynamically feasible plans in real time; (3) a multi-stage training strategy using supervised fine-tuning to elicit reasoning and reinforcement learning (RL) to optimize reasoning quality via large reasoning model feedback and enforce reasoning-action consistency. Evaluation shows AR1 achieves up to a 12% improvement in planning accuracy on challenging cases compared to a trajectory-only baseline, with a 35% reduction in off-road rate and 25% reduction in close encounter rate in closed-loop simulation. RL post-training improves reasoning quality by 45% as measured by a large reasoning model critic and reasoning-action consistency by 37%. Model scaling from 0.5B to 7B parameters shows consistent improvements. On-vehicle road tests confirm real-time performance (99 ms latency) and successful urban deployment. By bridging interpretable reasoning with precise control, AR1 demonstrates a practical path towards Level 4 autonomous driving. We plan to release AR1 models and a subset of the CoC in a future update.

Katie Z Luo, Xinshuo Weng, Yan Wang et al.

Autonomous driving has traditionally relied heavily on costly and labor-intensive High Definition (HD) maps, hindering scalability. In contrast, Standard Definition (SD) maps are more affordable and have worldwide coverage, offering a scalable alternative. In this work, we systematically explore the effect of SD maps for real-time lane-topology understanding. We propose a novel framework to integrate SD maps into online map prediction and propose a Transformer-based encoder, SD Map Encoder Representations from transFormers, to leverage priors in SD maps for the lane-topology prediction task. This enhancement consistently and significantly boosts (by up to 60%) lane detection and topology prediction on current state-of-the-art online map prediction methods without bells and whistles and can be immediately incorporated into any Transformer-based lane-topology method. Code is available at https://github.com/NVlabs/SMERF.

Aditi, Niket Agarwal, Arslan Ali et al.

We introduce Cosmos 3, a family of omnimodal world models designed to jointly process and generate language, image, video, audio, and action sequences within a unified mixture-of-transformers architecture. By supporting highly flexible input-output configurations, Cosmos 3 seamlessly unifies critical modalities for Physical AI -- effectively subsuming vision-language models, video generators, world simulators, and world-action models into a single framework. Our evaluation demonstrates that Cosmos 3 establishes a new state-of-the-art across a diverse suite of understanding and generation tasks, demonstrating omnimodal world models as scalable, general-purpose backbones for embodied agents. Our post-trained Cosmos 3 models were ranked as the best open-source Text-to-Image and Image-to-Video models by Artificial Analysis, and the best policy model by RoboArena at the time the technical report was written. To accelerate open research and deployment in Physical AI, we make our code, model checkpoints, curated synthetic datasets, and evaluation benchmark available under the Linux Foundation's OpenMDW-1.1 https://openmdw.ai/license/1-1/ License at https://github.com/nvidia/cosmos}{github.com/nvidia/cosmos and https://huggingface.co/collections/nvidia/cosmos3 . The project website is available at https://research.nvidia.com/labs/cosmos-lab/cosmos3 .

Shuhan Tan, Boris Ivanovic, Xinshuo Weng et al.

Simulation forms the backbone of modern self-driving development. Simulators help develop, test, and improve driving systems without putting humans, vehicles, or their environment at risk. However, simulators face a major challenge: They rely on realistic, scalable, yet interesting content. While recent advances in rendering and scene reconstruction make great strides in creating static scene assets, modeling their layout, dynamics, and behaviors remains challenging. In this work, we turn to language as a source of supervision for dynamic traffic scene generation. Our model, LCTGen, combines a large language model with a transformer-based decoder architecture that selects likely map locations from a dataset of maps, and produces an initial traffic distribution, as well as the dynamics of each vehicle. LCTGen outperforms prior work in both unconditional and conditional traffic scene generation in terms of realism and fidelity. Code and video will be available at https://ariostgx.github.io/lctgen.

Yulong Cao, Danfei Xu, Xinshuo Weng et al.

Trajectory prediction using deep neural networks (DNNs) is an essential component of autonomous driving (AD) systems. However, these methods are vulnerable to adversarial attacks, leading to serious consequences such as collisions. In this work, we identify two key ingredients to defend trajectory prediction models against adversarial attacks including (1) designing effective adversarial training methods and (2) adding domain-specific data augmentation to mitigate the performance degradation on clean data. We demonstrate that our method is able to improve the performance by 46% on adversarial data and at the cost of only 3% performance degradation on clean data, compared to the model trained with clean data. Additionally, compared to existing robust methods, our method can improve performance by 21% on adversarial examples and 9% on clean data. Our robust model is evaluated with a planner to study its downstream impacts. We demonstrate that our model can significantly reduce the severe accident rates (e.g., collisions and off-road driving).

Jiawei Yang, Boris Ivanovic, Or Litany et al.

We present EmerNeRF, a simple yet powerful approach for learning spatial-temporal representations of dynamic driving scenes. Grounded in neural fields, EmerNeRF simultaneously captures scene geometry, appearance, motion, and semantics via self-bootstrapping. EmerNeRF hinges upon two core components: First, it stratifies scenes into static and dynamic fields. This decomposition emerges purely from self-supervision, enabling our model to learn from general, in-the-wild data sources. Second, EmerNeRF parameterizes an induced flow field from the dynamic field and uses this flow field to further aggregate multi-frame features, amplifying the rendering precision of dynamic objects. Coupling these three fields (static, dynamic, and flow) enables EmerNeRF to represent highly-dynamic scenes self-sufficiently, without relying on ground truth object annotations or pre-trained models for dynamic object segmentation or optical flow estimation. Our method achieves state-of-the-art performance in sensor simulation, significantly outperforming previous methods when reconstructing static (+2.93 PSNR) and dynamic (+3.70 PSNR) scenes. In addition, to bolster EmerNeRF's semantic generalization, we lift 2D visual foundation model features into 4D space-time and address a general positional bias in modern Transformers, significantly boosting 3D perception performance (e.g., 37.50% relative improvement in occupancy prediction accuracy on average). Finally, we construct a diverse and challenging 120-sequence dataset to benchmark neural fields under extreme and highly-dynamic settings.

Ran Tian, Boyi Li, Xinshuo Weng et al.

The autonomous driving industry is increasingly adopting end-to-end learning from sensory inputs to minimize human biases in system design. Traditional end-to-end driving models, however, suffer from long-tail events due to rare or unseen inputs within their training distributions. To address this, we propose TOKEN, a novel Multi-Modal Large Language Model (MM-LLM) that tokenizes the world into object-level knowledge, enabling better utilization of LLM's reasoning capabilities to enhance autonomous vehicle planning in long-tail scenarios. TOKEN effectively alleviates data scarcity and inefficient tokenization by leveraging a traditional end-to-end driving model to produce condensed and semantically enriched representations of the scene, which are optimized for LLM planning compatibility through deliberate representation and reasoning alignment training stages. Our results demonstrate that TOKEN excels in grounding, reasoning, and planning capabilities, outperforming existing frameworks with a 27% reduction in trajectory L2 error and a 39% decrease in collision rates in long-tail scenarios. Additionally, our work highlights the importance of representation alignment and structured reasoning in sparking the common-sense reasoning capabilities of MM-LLMs for effective planning.

Shuhan Tan, Boris Ivanovic, Yuxiao Chen et al.

Simulation stands as a cornerstone for safe and efficient autonomous driving development. At its core a simulation system ought to produce realistic, reactive, and controllable traffic patterns. In this paper, we propose ProSim, a multimodal promptable closed-loop traffic simulation framework. ProSim allows the user to give a complex set of numerical, categorical or textual prompts to instruct each agent's behavior and intention. ProSim then rolls out a traffic scenario in a closed-loop manner, modeling each agent's interaction with other traffic participants. Our experiments show that ProSim achieves high prompt controllability given different user prompts, while reaching competitive performance on the Waymo Sim Agents Challenge when no prompt is given. To support research on promptable traffic simulation, we create ProSim-Instruct-520k, a multimodal prompt-scenario paired driving dataset with over 10M text prompts for over 520k real-world driving scenarios. We will release code of ProSim as well as data and labeling tools of ProSim-Instruct-520k at https://ariostgx.github.io/ProSim.

Boyi Li, Ligeng Zhu, Ran Tian et al.

We propose Wolf, a WOrLd summarization Framework for accurate video captioning. Wolf is an automated captioning framework that adopts a mixture-of-experts approach, leveraging complementary strengths of Vision Language Models (VLMs). By utilizing both image and video models, our framework captures different levels of information and summarizes them efficiently. Our approach can be applied to enhance video understanding, auto-labeling, and captioning. To evaluate caption quality, we introduce CapScore, an LLM-based metric to assess the similarity and quality of generated captions compared to the ground truth captions. We further build four human-annotated datasets in three domains: autonomous driving, general scenes, and robotics, to facilitate comprehensive comparisons. We show that Wolf achieves superior captioning performance compared to state-of-the-art approaches from the research community (VILA1.5, CogAgent) and commercial solutions (Gemini-Pro-1.5, GPT-4V). For instance, in comparison with GPT-4V, Wolf improves CapScore both quality-wise by 55.6% and similarity-wise by 77.4% on challenging driving videos. Finally, we establish a benchmark for video captioning and introduce a leaderboard, aiming to accelerate advancements in video understanding, captioning, and data alignment. Webpage: https://wolfv0.github.io/.

Daniel Dauner, Marcel Hallgarten, Tianyu Li et al.

Benchmarking vision-based driving policies is challenging. On one hand, open-loop evaluation with real data is easy, but these results do not reflect closed-loop performance. On the other, closed-loop evaluation is possible in simulation, but is hard to scale due to its significant computational demands. Further, the simulators available today exhibit a large domain gap to real data. This has resulted in an inability to draw clear conclusions from the rapidly growing body of research on end-to-end autonomous driving. In this paper, we present NAVSIM, a middle ground between these evaluation paradigms, where we use large datasets in combination with a non-reactive simulator to enable large-scale real-world benchmarking. Specifically, we gather simulation-based metrics, such as progress and time to collision, by unrolling bird's eye view abstractions of the test scenes for a short simulation horizon. Our simulation is non-reactive, i.e., the evaluated policy and environment do not influence each other. As we demonstrate empirically, this decoupling allows open-loop metric computation while being better aligned with closed-loop evaluations than traditional displacement errors. NAVSIM enabled a new competition held at CVPR 2024, where 143 teams submitted 463 entries, resulting in several new insights. On a large set of challenging scenarios, we observe that simple methods with moderate compute requirements such as TransFuser can match recent large-scale end-to-end driving architectures such as UniAD. Our modular framework can potentially be extended with new datasets, data curation strategies, and metrics, and will be continually maintained to host future challenges. Our code is available at https://github.com/autonomousvision/navsim.

Jinhyung Park, Xinshuo Weng, Yunze Man et al.

Point clouds and RGB images are naturally complementary modalities for 3D visual understanding - the former provides sparse but accurate locations of points on objects, while the latter contains dense color and texture information. Despite this potential for close sensor fusion, many methods train two models in isolation and use simple feature concatenation to represent 3D sensor data. This separated training scheme results in potentially sub-optimal performance and prevents 3D tasks from being used to benefit 2D tasks that are often useful on their own. To provide a more integrated approach, we propose a novel Multi-Modality Task Cascade network (MTC-RCNN) that leverages 3D box proposals to improve 2D segmentation predictions, which are then used to further refine the 3D boxes. We show that including a 2D network between two stages of 3D modules significantly improves both 2D and 3D task performance. Moreover, to prevent the 3D module from over-relying on the overfitted 2D predictions, we propose a dual-head 2D segmentation training and inference scheme, allowing the 2nd 3D module to learn to interpret imperfect 2D segmentation predictions. Evaluating our model on the challenging SUN RGB-D dataset, we improve upon state-of-the-art results of both single modality and fusion networks by a large margin ($\textbf{+3.8}$ mAP@0.5). Code will be released $\href{https://github.com/Divadi/MTC_RCNN}{\text{here.}}$

Xuanyi Dong, Yi Yang, Shih-En Wei et al.

We present Supervision by Registration and Triangulation (SRT), an unsupervised approach that utilizes unlabeled multi-view video to improve the accuracy and precision of landmark detectors. Being able to utilize unlabeled data enables our detectors to learn from massive amounts of unlabeled data freely available and not be limited by the quality and quantity of manual human annotations. To utilize unlabeled data, there are two key observations: (1) the detections of the same landmark in adjacent frames should be coherent with registration, i.e., optical flow. (2) the detections of the same landmark in multiple synchronized and geometrically calibrated views should correspond to a single 3D point, i.e., multi-view consistency. Registration and multi-view consistency are sources of supervision that do not require manual labeling, thus it can be leveraged to augment existing training data during detector training. End-to-end training is made possible by differentiable registration and 3D triangulation modules. Experiments with 11 datasets and a newly proposed metric to measure precision demonstrate accuracy and precision improvements in landmark detection on both images and video. Code is available at https://github.com/D-X-Y/landmark-detection.

Yongxin Wang, Kris Kitani, Xinshuo Weng

Object detection and data association are critical components in multi-object tracking (MOT) systems. Despite the fact that the two components are dependent on each other, prior works often design detection and data association modules separately which are trained with separate objectives. As a result, one cannot back-propagate the gradients and optimize the entire MOT system, which leads to sub-optimal performance. To address this issue, recent works simultaneously optimize detection and data association modules under a joint MOT framework, which has shown improved performance in both modules. In this work, we propose a new instance of joint MOT approach based on Graph Neural Networks (GNNs). The key idea is that GNNs can model relations between variable-sized objects in both the spatial and temporal domains, which is essential for learning discriminative features for detection and data association. Through extensive experiments on the MOT15/16/17/20 datasets, we demonstrate the effectiveness of our GNN-based joint MOT approach and show state-of-the-art performance for both detection and MOT tasks. Our code is available at: https://github.com/yongxinw/GSDT

Xinshuo Weng, Yongxin Wang, Yunze Man et al.

3D Multi-object tracking (MOT) is crucial to autonomous systems. Recent work uses a standard tracking-by-detection pipeline, where feature extraction is first performed independently for each object in order to compute an affinity matrix. Then the affinity matrix is passed to the Hungarian algorithm for data association. A key process of this standard pipeline is to learn discriminative features for different objects in order to reduce confusion during data association. In this work, we propose two techniques to improve the discriminative feature learning for MOT: (1) instead of obtaining features for each object independently, we propose a novel feature interaction mechanism by introducing the Graph Neural Network. As a result, the feature of one object is informed of the features of other objects so that the object feature can lean towards the object with similar feature (i.e., object probably with a same ID) and deviate from objects with dissimilar features (i.e., object probably with different IDs), leading to a more discriminative feature for each object; (2) instead of obtaining the feature from either 2D or 3D space in prior work, we propose a novel joint feature extractor to learn appearance and motion features from 2D and 3D space simultaneously. As features from different modalities often have complementary information, the joint feature can be more discriminate than feature from each individual modality. To ensure that the joint feature extractor does not heavily rely on one modality, we also propose an ensemble training paradigm. Through extensive evaluation, our proposed method achieves state-of-the-art performance on KITTI and nuScenes 3D MOT benchmarks. Our code will be made available at https://github.com/xinshuoweng/GNN3DMOT

Xinshuo Weng, Jianren Wang, David Held et al.

3D multi-object tracking (MOT) is an essential component for many applications such as autonomous driving and assistive robotics. Recent work on 3D MOT focuses on developing accurate systems giving less attention to practical considerations such as computational cost and system complexity. In contrast, this work proposes a simple real-time 3D MOT system. Our system first obtains 3D detections from a LiDAR point cloud. Then, a straightforward combination of a 3D Kalman filter and the Hungarian algorithm is used for state estimation and data association. Additionally, 3D MOT datasets such as KITTI evaluate MOT methods in the 2D space and standardized 3D MOT evaluation tools are missing for a fair comparison of 3D MOT methods. Therefore, we propose a new 3D MOT evaluation tool along with three new metrics to comprehensively evaluate 3D MOT methods. We show that, although our system employs a combination of classical MOT modules, we achieve state-of-the-art 3D MOT performance on two 3D MOT benchmarks (KITTI and nuScenes). Surprisingly, although our system does not use any 2D data as inputs, we achieve competitive performance on the KITTI 2D MOT leaderboard. Our proposed system runs at a rate of $207.4$ FPS on the KITTI dataset, achieving the fastest speed among all modern MOT systems. To encourage standardized 3D MOT evaluation, our system and evaluation code are made publicly available at https://github.com/xinshuoweng/AB3DMOT.

Xinshuo Weng, Kris Kitani

Monocular 3D scene understanding tasks, such as object size estimation, heading angle estimation and 3D localization, is challenging. Successful modern day methods for 3D scene understanding require the use of a 3D sensor. On the other hand, single image based methods have significantly worse performance. In this work, we aim at bridging the performance gap between 3D sensing and 2D sensing for 3D object detection by enhancing LiDAR-based algorithms to work with single image input. Specifically, we perform monocular depth estimation and lift the input image to a point cloud representation, which we call pseudo-LiDAR point cloud. Then we can train a LiDAR-based 3D detection network with our pseudo-LiDAR end-to-end. Following the pipeline of two-stage 3D detection algorithms, we detect 2D object proposals in the input image and extract a point cloud frustum from the pseudo-LiDAR for each proposal. Then an oriented 3D bounding box is detected for each frustum. To handle the large amount of noise in the pseudo-LiDAR, we propose two innovations: (1) use a 2D-3D bounding box consistency constraint, adjusting the predicted 3D bounding box to have a high overlap with its corresponding 2D proposal after projecting onto the image; (2) use the instance mask instead of the bounding box as the representation of 2D proposals, in order to reduce the number of points not belonging to the object in the point cloud frustum. Through our evaluation on the KITTI benchmark, we achieve the top-ranked performance on both bird's eye view and 3D object detection among all monocular methods, effectively quadrupling the performance over previous state-of-the-art. Our code is available at https://github.com/xinshuoweng/Mono3D_PLiDAR.

Jang Hyun Cho, Boris Ivanovic, Yulong Cao et al.

Multi-modal large language models (MLLMs) have shown incredible capabilities in a variety of 2D vision and language tasks. We extend MLLMs' perceptual capabilities to ground and reason about images in 3-dimensional space. To that end, we first develop a large-scale pre-training dataset for 2D and 3D called LV3D by combining multiple existing 2D and 3D recognition datasets under a common task formulation: as multi-turn question-answering. Next, we introduce a new MLLM named Cube-LLM and pre-train it on LV3D. We show that pure data scaling makes a strong 3D perception capability without 3D specific architectural design or training objective. Cube-LLM exhibits intriguing properties similar to LLMs: (1) Cube-LLM can apply chain-of-thought prompting to improve 3D understanding from 2D context information. (2) Cube-LLM can follow complex and diverse instructions and adapt to versatile input and output formats. (3) Cube-LLM can be visually prompted such as 2D box or a set of candidate 3D boxes from specialists. Our experiments on outdoor benchmarks demonstrate that Cube-LLM significantly outperforms existing baselines by 21.3 points of AP-BEV on the Talk2Car dataset for 3D grounded reasoning and 17.7 points on the DriveLM dataset for complex reasoning about driving scenarios, respectively. Cube-LLM also shows competitive results in general MLLM benchmarks such as refCOCO for 2D grounding with (87.0) average score, as well as visual question answering benchmarks such as VQAv2, GQA, SQA, POPE, etc. for complex reasoning. Our project is available at https://janghyuncho.github.io/Cube-LLM.

Xinshuo Weng, Boris Ivanovic, Marco Pavone

Recently, there has been tremendous progress in developing each individual module of the standard perception-planning robot autonomy pipeline, including detection, tracking, prediction of other agents' trajectories, and ego-agent trajectory planning. Nevertheless, there has been less attention given to the principled integration of these components, particularly in terms of the characterization and mitigation of cascading errors. This paper addresses the problem of cascading errors by focusing on the coupling between the tracking and prediction modules. First, by using state-of-the-art tracking and prediction tools, we conduct a comprehensive experimental evaluation of how severely errors stemming from tracking can impact prediction performance. On the KITTI and nuScenes datasets, we find that predictions consuming tracked trajectories as inputs (the typical case in practice) can experience a significant (even order of magnitude) drop in performance in comparison to the idealized setting where ground truth past trajectories are used as inputs. To address this issue, we propose a multi-hypothesis tracking and prediction framework. Rather than relying on a single set of tracking results for prediction, our framework simultaneously reasons about multiple sets of tracking results, thereby increasing the likelihood of including accurate tracking results as inputs to prediction. We show that this framework improves overall prediction performance over the standard single-hypothesis tracking-prediction pipeline by up to 34.2% on the nuScenes dataset, with even more significant improvements (up to ~70%) when restricting the evaluation to challenging scenarios involving identity switches and fragments -- all with an acceptable computation overhead.

Yunze Man, Xinshuo Weng, Prasanna Kumar Sivakuma et al.

LiDAR sensors can be used to obtain a wide range of measurement signals other than a simple 3D point cloud, and those signals can be leveraged to improve perception tasks like 3D object detection. A single laser pulse can be partially reflected by multiple objects along its path, resulting in multiple measurements called echoes. Multi-echo measurement can provide information about object contours and semi-transparent surfaces which can be used to better identify and locate objects. LiDAR can also measure surface reflectance (intensity of laser pulse return), as well as ambient light of the scene (sunlight reflected by objects). These signals are already available in commercial LiDAR devices but have not been used in most LiDAR-based detection models. We present a 3D object detection model which leverages the full spectrum of measurement signals provided by LiDAR. First, we propose a multi-signal fusion (MSF) module to combine (1) the reflectance and ambient features extracted with a 2D CNN, and (2) point cloud features extracted using a 3D graph neural network (GNN). Second, we propose a multi-echo aggregation (MEA) module to combine the information encoded in different set of echo points. Compared with traditional single echo point cloud methods, our proposed Multi-Signal LiDAR Detector (MSLiD) extracts richer context information from a wider range of sensing measurements and achieves more accurate 3D object detection. Experiments show that by incorporating the multi-modality of LiDAR, our method outperforms the state-of-the-art by up to 9.1%.

Yan Xu, Yu-Jhe Li, Xinshuo Weng et al.

We address the problem of estimating the 3D pose of a network of cameras for large-environment wide-baseline scenarios, e.g., cameras for construction sites, sports stadiums, and public spaces. This task is challenging since detecting and matching the same 3D keypoint observed from two very different camera views is difficult, making standard structure-from-motion (SfM) pipelines inapplicable. In such circumstances, treating people in the scene as "keypoints" and associating them across different camera views can be an alternative method for obtaining correspondences. Based on this intuition, we propose a method that uses ideas from person re-identification (re-ID) for wide-baseline camera calibration. Our method first employs a re-ID method to associate human bounding boxes across cameras, then converts bounding box correspondences to point correspondences, and finally solves for camera pose using multi-view geometry and bundle adjustment. Since our method does not require specialized calibration targets except for visible people, it applies to situations where frequent calibration updates are required. We perform extensive experiments on datasets captured from scenes of different sizes, camera settings (indoor and outdoor), and human activities (walking, playing basketball, construction). Experiment results show that our method achieves similar performance to standard SfM methods relying on manually labeled point correspondences.

Ye Yuan, Xinshuo Weng, Yanglan Ou et al.

Predicting accurate future trajectories of multiple agents is essential for autonomous systems, but is challenging due to the complex agent interaction and the uncertainty in each agent's future behavior. Forecasting multi-agent trajectories requires modeling two key dimensions: (1) time dimension, where we model the influence of past agent states over future states; (2) social dimension, where we model how the state of each agent affects others. Most prior methods model these two dimensions separately, e.g., first using a temporal model to summarize features over time for each agent independently and then modeling the interaction of the summarized features with a social model. This approach is suboptimal since independent feature encoding over either the time or social dimension can result in a loss of information. Instead, we would prefer a method that allows an agent's state at one time to directly affect another agent's state at a future time. To this end, we propose a new Transformer, AgentFormer, that jointly models the time and social dimensions. The model leverages a sequence representation of multi-agent trajectories by flattening trajectory features across time and agents. Since standard attention operations disregard the agent identity of each element in the sequence, AgentFormer uses a novel agent-aware attention mechanism that preserves agent identities by attending to elements of the same agent differently than elements of other agents. Based on AgentFormer, we propose a stochastic multi-agent trajectory prediction model that can attend to features of any agent at any previous timestep when inferring an agent's future position. The latent intent of all agents is also jointly modeled, allowing the stochasticity in one agent's behavior to affect other agents. Our method substantially improves the state of the art on well-established pedestrian and autonomous driving datasets.

Xinshuo Weng, Kris Kitani

3D multi-object tracking is an important component in robotic perception systems such as self-driving vehicles. Recent work follows a tracking-by-detection pipeline, which aims to match past tracklets with detections in the current frame. To avoid matching with false positive detections, prior work filters out detections with low confidence scores via a threshold. However, finding a proper threshold is non-trivial, which requires extensive manual search via ablation study. Also, this threshold is sensitive to many factors such as target object category so we need to re-search the threshold if these factors change. To ease this process, we propose to automatically select high-quality detections and remove the efforts needed for manual threshold search. Also, prior work often uses a single threshold per data sequence, which is sub-optimal in particular frames or for certain objects. Instead, we dynamically search threshold per frame or per object to further boost performance. Through experiments on KITTI and nuScenes, our method can filter out $45.7\%$ false positives while maintaining the recall, achieving new S.O.T.A. performance and removing the need for manually threshold tuning.

Xinshuo Weng, Ye Yuan, Kris Kitani

3D multi-object tracking (MOT) and trajectory forecasting are two critical components in modern 3D perception systems. We hypothesize that it is beneficial to unify both tasks under one framework to learn a shared feature representation of agent interaction. To evaluate this hypothesis, we propose a unified solution for 3D MOT and trajectory forecasting which also incorporates two additional novel computational units. First, we employ a feature interaction technique by introducing Graph Neural Networks (GNNs) to capture the way in which multiple agents interact with one another. The GNN is able to model complex hierarchical interactions, improve the discriminative feature learning for MOT association, and provide socially-aware context for trajectory forecasting. Second, we use a diversity sampling function to improve the quality and diversity of our forecasted trajectories. The learned sampling function is trained to efficiently extract a variety of outcomes from a generative trajectory distribution and helps avoid the problem of generating many duplicate trajectory samples. We show that our method achieves state-of-the-art performance on the KITTI dataset. Our project website is at http://www.xinshuoweng.com/projects/GNNTrkForecast.

Xinshuo Weng, Yongxin Wang, Yunze Man et al.

3D Multi-object tracking (MOT) is crucial to autonomous systems. Recent work often uses a tracking-by-detection pipeline, where the feature of each object is extracted independently to compute an affinity matrix. Then, the affinity matrix is passed to the Hungarian algorithm for data association. A key process of this pipeline is to learn discriminative features for different objects in order to reduce confusion during data association. To that end, we propose two innovative techniques: (1) instead of obtaining the features for each object independently, we propose a novel feature interaction mechanism by introducing Graph Neural Networks; (2) instead of obtaining the features from either 2D or 3D space as in prior work, we propose a novel joint feature extractor to learn appearance and motion features from 2D and 3D space. Through experiments on the KITTI dataset, our proposed method achieves state-of-the-art 3D MOT performance. Our project website is at http://www.xinshuoweng.com/projects/GNN3DMOT.

Xinshuo Weng, Jianren Wang, David Held et al.

3D multi-object tracking (MOT) is essential to applications such as autonomous driving. Recent work focuses on developing accurate systems giving less attention to computational cost and system complexity. In contrast, this work proposes a simple real-time 3D MOT system with strong performance. Our system first obtains 3D detections from a LiDAR point cloud. Then, a straightforward combination of a 3D Kalman filter and the Hungarian algorithm is used for state estimation and data association. Additionally, 3D MOT datasets such as KITTI evaluate MOT methods in 2D space and standardized 3D MOT evaluation tools are missing for a fair comparison of 3D MOT methods. We propose a new 3D MOT evaluation tool along with three new metrics to comprehensively evaluate 3D MOT methods. We show that, our proposed method achieves strong 3D MOT performance on KITTI and runs at a rate of $207.4$ FPS on the KITTI dataset, achieving the fastest speed among modern 3D MOT systems. Our code is publicly available at http://www.xinshuoweng.com/projects/AB3DMOT.

Xi Sun, Xinshuo Weng, Kris Kitani

We aim to enable robots to visually localize a target person through the aid of an additional sensing modality -- the target person's 3D inertial measurements. The need for such technology may arise when a robot is to meet person in a crowd for the first time or when an autonomous vehicle must rendezvous with a rider amongst a crowd without knowing the appearance of the person in advance. A person's inertial information can be measured with a wearable device such as a smart-phone and can be shared selectively with an autonomous system during the rendezvous. We propose a method to learn a visual-inertial feature space in which the motion of a person in video can be easily matched to the motion measured by a wearable inertial measurement unit (IMU). The transformation of the two modalities into the joint feature space is learned through the use of a contrastive loss which forces inertial motion features and video motion features generated by the same person to lie close in the joint feature space. To validate our approach, we compose a dataset of over 60,000 video segments of moving people along with wearable IMU data. Our experiments show that our proposed method is able to accurately localize a target person with 80.7% accuracy using only 5 seconds of IMU data and video.

Xinshuo Weng, Jianren Wang, Sergey Levine et al.

Many autonomous systems forecast aspects of the future in order to aid decision-making. For example, self-driving vehicles and robotic manipulation systems often forecast future object poses by first detecting and tracking objects. However, this detect-then-forecast pipeline is expensive to scale, as pose forecasting algorithms typically require labeled sequences of object poses, which are costly to obtain in 3D space. Can we scale performance without requiring additional labels? We hypothesize yes, and propose inverting the detect-then-forecast pipeline. Instead of detecting, tracking and then forecasting the objects, we propose to first forecast 3D sensor data (e.g., point clouds with $100$k points) and then detect/track objects on the predicted point cloud sequences to obtain future poses, i.e., a forecast-then-detect pipeline. This inversion makes it less expensive to scale pose forecasting, as the sensor data forecasting task requires no labels. Part of this work's focus is on the challenging first step -- Sequential Pointcloud Forecasting (SPF), for which we also propose an effective approach, SPFNet. To compare our forecast-then-detect pipeline relative to the detect-then-forecast pipeline, we propose an evaluation procedure and two metrics. Through experiments on a robotic manipulation dataset and two driving datasets, we show that SPFNet is effective for the SPF task, our forecast-then-detect pipeline outperforms the detect-then-forecast approaches to which we compared, and that pose forecasting performance improves with the addition of unlabeled data.

Xinshuo Weng, Ye Yuan, Kris Kitani

Multi-object tracking (MOT) and trajectory prediction are two critical components in modern 3D perception systems that require accurate modeling of multi-agent interaction. We hypothesize that it is beneficial to unify both tasks under one framework in order to learn a shared feature representation of agent interaction. Furthermore, instead of performing tracking and prediction sequentially which can propagate errors from tracking to prediction, we propose a parallelized framework to mitigate the issue. Also, our parallel track-forecast framework incorporates two additional novel computational units. First, we use a feature interaction technique by introducing Graph Neural Networks (GNNs) to capture the way in which agents interact with one another. The GNN is able to improve discriminative feature learning for MOT association and provide socially-aware contexts for trajectory prediction. Second, we use a diversity sampling function to improve the quality and diversity of our forecasted trajectories. The learned sampling function is trained to efficiently extract a variety of outcomes from a generative trajectory distribution and helps avoid the problem of generating duplicate trajectory samples. We evaluate on KITTI and nuScenes datasets showing that our method with socially-aware feature learning and diversity sampling achieves new state-of-the-art performance on 3D MOT and trajectory prediction. Project website is: https://www.xinshuoweng.com/projects/PTP

Yu-Jhe Li, Zhengyi Luo, Xinshuo Weng et al.

Person re-identification (re-ID) aims to recognize instances of the same person contained in multiple images taken across different cameras. Existing methods for re-ID tend to rely heavily on the assumption that both query and gallery images of the same person have the same clothing. Unfortunately, this assumption may not hold for datasets captured over long periods of time (e.g., weeks, months or years). To tackle the re-ID problem in the context of clothing changes, we propose a novel representation learning model which is able to generate a body shape feature representation without being affected by clothing color or patterns. We call our model the Color Agnostic Shape Extraction Network (CASE-Net). CASE-Net learns a representation of identity that depends only on body shape via adversarial learning and feature disentanglement. Due to the lack of large-scale re-ID datasets which contain clothing changes for the same person, we propose two synthetic datasets for evaluation. We create a rendered dataset SMPL-reID with different clothes patterns and a synthesized dataset Div-Market with different clothing color to simulate two types of clothing changes. The quantitative and qualitative results across 5 datasets (SMPL-reID, Div-Market, two benchmark re-ID datasets, a cross-modality re-ID dataset) confirm the robustness and superiority of our approach against several state-of-the-art approaches

Xinshuo Weng, Kris Kitani

We focus on the word-level visual lipreading, which requires recognizing the word being spoken, given only the video but not the audio. State-of-the-art methods explore the use of end-to-end neural networks, including a shallow (up to three layers) 3D convolutional neural network (CNN) + a deep 2D CNN (e.g., ResNet) as the front-end to extract visual features, and a recurrent neural network (e.g., bidirectional LSTM) as the back-end for classification. In this work, we propose to replace the shallow 3D CNNs + deep 2D CNNs front-end with recent successful deep 3D CNNs --- two-stream (i.e., grayscale video and optical flow streams) I3D. We evaluate different combinations of front-end and back-end modules with the grayscale video and optical flow inputs on the LRW dataset. The experiments show that, compared to the shallow 3D CNNs + deep 2D CNNs front-end, the deep 3D CNNs front-end with pre-training on the large-scale image and video datasets (e.g., ImageNet and Kinetics) can improve the classification accuracy. Also, we demonstrate that using the optical flow input alone can achieve comparable performance as using the grayscale video as input. Moreover, the two-stream network using both the grayscale video and optical flow inputs can further improve the performance. Overall, our two-stream I3D front-end with a Bi-LSTM back-end results in an absolute improvement of 5.3% over the previous art on the LRW dataset.

Xinshuo Weng

We focus on the word-level visual lipreading, which requires to decode the word from the speaker's video. Recently, many state-of-the-art visual lipreading methods explore the end-to-end trainable deep models, involving the use of 2D convolutional networks (e.g., ResNet) as the front-end visual feature extractor and the sequential model (e.g., Bi-LSTM or Bi-GRU) as the back-end. Although a deep 2D convolution neural network can provide informative image-based features, it ignores the temporal motion existing between the adjacent frames. In this work, we investigate the spatial-temporal capacity power of I3D (Inflated 3D ConvNet) for visual lipreading. We demonstrate that, after pre-trained on the large-scale video action recognition dataset (e.g., Kinetics), our models show a considerable improvement of performance on the task of lipreading. A comparison between a set of video model architectures and input data representation is also reported. Our extensive experiments on LRW shows that a two-stream I3D model with RGB video and optical flow as the inputs achieves the state-of-the-art performance.

Aashi Manglik, Xinshuo Weng, Eshed Ohn-Bar et al.

We explore the possibility of using a single monocular camera to forecast the time to collision between a suitcase-shaped robot being pushed by its user and other nearby pedestrians. We develop a purely image-based deep learning approach that directly estimates the time to collision without the need of relying on explicit geometric depth estimates or velocity information to predict future collisions. While previous work has focused on detecting immediate collision in the context of navigating Unmanned Aerial Vehicles, the detection was limited to a binary variable (i.e., collision or no collision). We propose a more fine-grained approach to collision forecasting by predicting the exact time to collision in terms of milliseconds, which is more helpful for collision avoidance in the context of dynamic path planning. To evaluate our method, we have collected a novel dataset of over 13,000 indoor video segments each showing a trajectory of at least one person ending in a close proximity (a near collision) with the camera mounted on a mobile suitcase-shaped platform. Using this dataset, we do extensive experimentation on different temporal windows as input using an exhaustive list of state-of-the-art convolutional neural networks (CNNs). Our results show that our proposed multi-stream CNN is the best model for predicting time to near-collision. The average prediction error of our time to near collision is 0.75 seconds across the test videos.

Sen Wang, Daoyuan Jia, Xinshuo Weng

Reinforcement learning has steadily improved and outperform human in lots of traditional games since the resurgence of deep neural network. However, these success is not easy to be copied to autonomous driving because the state spaces in real world are extreme complex and action spaces are continuous and fine control is required. Moreover, the autonomous driving vehicles must also keep functional safety under the complex environments. To deal with these challenges, we first adopt the deep deterministic policy gradient (DDPG) algorithm, which has the capacity to handle complex state and action spaces in continuous domain. We then choose The Open Racing Car Simulator (TORCS) as our environment to avoid physical damage. Meanwhile, we select a set of appropriate sensor information from TORCS and design our own rewarder. In order to fit DDPG algorithm to TORCS, we design our network architecture for both actor and critic inside DDPG paradigm. To demonstrate the effectiveness of our model, We evaluate on different modes in TORCS and show both quantitative and qualitative results.

Xinshuo Weng, Wentao Han

Across a majority of modern learning-based tracking systems, expensive annotations are needed to achieve state-of-the-art performance. In contrast, the Lucas-Kanade (LK) algorithm works well without any annotation. However, LK has a strong assumption of photometric (brightness) consistency on image intensity and is easy to drift because of large motion, occlusion, and aperture problem. To relax the assumption and alleviate the drift problem, we propose CyLKs, a data-driven way of training Lucas-Kanade in an unsupervised manner. CyLKs learns a feature transformation through CNNs, transforming the input images to a feature space which is especially favorable to LK tracking. During training, we perform differentiable Lucas-Kanade forward and backward on the convolutional feature maps, and then minimize the re-projection error. During testing, we perform the LK tracking on the learned features. We apply our model to the task of landmark tracking and perform experiments on datasets of THUMOS and 300VW.

Shangxuan Wu, Xinshuo Weng

Most existing methods for object segmentation in computer vision are formulated as a labeling task. This, in general, could be transferred to a pixel-wise label assignment task, which is quite similar to the structure of hidden Markov random field. In terms of Markov random field, each pixel can be regarded as a state and has a transition probability to its neighbor pixel, the label behind each pixel is a latent variable and has an emission probability from its corresponding state. In this paper, we reviewed several modern image labeling methods based on Markov random field and conditional random Field. And we compare the result of these methods with some classical image labeling methods. The experiment demonstrates that the introduction of Markov random field and conditional random field make a big difference in the segmentation result.

Yunze Man, Xinshuo Weng, Xi Li et al.

We focus on estimating the 3D orientation of the ground plane from a single image. We formulate the problem as an inter-mingled multi-task prediction problem by jointly optimizing for pixel-wise surface normal direction, ground plane segmentation, and depth estimates. Specifically, our proposed model, GroundNet, first estimates the depth and surface normal in two separate streams, from which two ground plane normals are then computed deterministically. To leverage the geometric correlation between depth and normal, we propose to add a consistency loss on top of the computed ground plane normals. In addition, a ground segmentation stream is used to isolate the ground regions so that we can selectively back-propagate parameter updates through only the ground regions in the image. Our method achieves the top-ranked performance on ground plane normal estimation and horizon line detection on the real-world outdoor datasets of ApolloScape and KITTI, improving the performance of previous art by up to 17.7% relatively.

Xuanyi Dong, Shoou-I Yu, Xinshuo Weng et al.

In this paper, we present supervision-by-registration, an unsupervised approach to improve the precision of facial landmark detectors on both images and video. Our key observation is that the detections of the same landmark in adjacent frames should be coherent with registration, i.e., optical flow. Interestingly, the coherency of optical flow is a source of supervision that does not require manual labeling, and can be leveraged during detector training. For example, we can enforce in the training loss function that a detected landmark at frame$_{t-1}$ followed by optical flow tracking from frame$_{t-1}$ to frame$_t$ should coincide with the location of the detection at frame$_t$. Essentially, supervision-by-registration augments the training loss function with a registration loss, thus training the detector to have output that is not only close to the annotations in labeled images, but also consistent with registration on large amounts of unlabeled videos. End-to-end training with the registration loss is made possible by a differentiable Lucas-Kanade operation, which computes optical flow registration in the forward pass, and back-propagates gradients that encourage temporal coherency in the detector. The output of our method is a more precise image-based facial landmark detector, which can be applied to single images or video. With supervision-by-registration, we demonstrate (1) improvements in facial landmark detection on both images (300W, ALFW) and video (300VW, Youtube-Celebrities), and (2) significant reduction of jittering in video detections.

Xinshuo Weng, Shangxuan Wu, Fares Beainy et al.

Across a majority of pedestrian detection datasets, it is typically assumed that pedestrians will be standing upright with respect to the image coordinate system. This assumption, however, is not always valid for many vision-equipped mobile platforms such as mobile phones, UAVs or construction vehicles on rugged terrain. In these situations, the motion of the camera can cause images of pedestrians to be captured at extreme angles. This can lead to very poor pedestrian detection performance when using standard pedestrian detectors. To address this issue, we propose a Rotational Rectification Network (R2N) that can be inserted into any CNN-based pedestrian (or object) detector to adapt it to significant changes in camera rotation. The rotational rectification network uses a 2D rotation estimation module that passes rotational information to a spatial transformer network to undistort image features. To enable robust rotation estimation, we propose a Global Polar Pooling (GP-Pooling) operator to capture rotational shifts in convolutional features. Through our experiments, we show how our rotational rectification network can be used to improve the performance of the state-of-the-art pedestrian detector under heavy image rotation by up to 45%

Namhoon Lee, Xinshuo Weng, Vishnu Naresh Boddeti et al.

We introduce the concept of a Visual Compiler that generates a scene specific pedestrian detector and pose estimator without any pedestrian observations. Given a single image and auxiliary scene information in the form of camera parameters and geometric layout of the scene, the Visual Compiler first infers geometrically and photometrically accurate images of humans in that scene through the use of computer graphics rendering. Using these renders we learn a scene-and-region specific spatially-varying fully convolutional neural network, for simultaneous detection, pose estimation and segmentation of pedestrians. We demonstrate that when real human annotated data is scarce or non-existent, our data generation strategy can provide an excellent solution for bootstrapping human detection and pose estimation. Experimental results show that our approach outperforms off-the-shelf state-of-the-art pedestrian detectors and pose estimators that are trained on real data.