Estimating 3D human poses from 2D images remains challenging due to occlusions and projective ambiguity. Multi-view learning-based approaches mitigate these issues but often fail to generalize to real-world scenarios, as large-scale multi-view datasets with 3D ground truth are scarce and captured under constrained conditions. To overcome this limitation, recent methods rely on 2D pose estimation combined with 2D-to-3D pose lifting trained on synthetic data. Building on our previous MPL framework, we propose RUMPL, a transformer-based 3D pose lifter that introduces a 3D ray-based representation of 2D keypoints. This formulation enables a model agnostic to camera parameters that can be universally deployed across arbitrary camera configurations in a given area without retraining or fine-tuning. A new View Fusion Transformer leverages learned fused-ray tokens to aggregate information along rays, further improving multi-view consistency. Evaluation on standard benchmarks shows that RUMPL significantly outperforms existing methods, yielding a 56.6% MPJPE (All KP) reduction on Human3.6M over triangulation-based methods and exceeding 70% improvement on the CMU Panoptic dataset when compared to transformer-based image-representation approaches. Results on new benchmarks, including in-the-wild multi-view and multi-person datasets, confirm its robustness and scalability.

Multi-modal vehicle trajectory prediction is crucial for autonomous driving in dynamic environments. Despite the significant progress in the field, the uncertainty and heterogeneity caused by the diversity in driving intentions and driving scenes still present major challenges to multi-modal prediction. Existing query-based prediction paradigms take into account the social context arising from the driving scene, but neglect priors regarding vehicle intentions that the historical trajectory may contain. We propose a Social-aware Water-wave Information Transmission Recurrent Acceleration Network, abbreviated as S-WITRAN, based on decoupling multi-modal prediction into an ego-aware and a social-aware learning stage. The ego-aware stage aims to relax the constraints from the driving scene to explore a diversity of future trajectory candidates. The social-aware stage aligns the candidates with respect to the social context arising from the driving scene. The information transmission is designed to extract from the vehicle’s historical trajectory priors about its possible intentions and dynamic states, which are integrated to form a multi-modal set of trajectories. Extensive experiments on the NGSIM, highD, and Argoverse2 datasets, as well as a benchmark evaluation in UniTraj, demonstrate the superior performance of the proposed model in both map-free and map-based datasets.

Conditioning image generation facilitates seamless editing and the creation of photorealistic images. However, conditioning on noisy or Out-of-Distribution (OoD) images poses significant challenges, particularly in balancing fidelity to the input and realism of the output. We introduce Confident Ordinary Differential Editing (CODE), a novel approach for image synthesis that effectively handles OoD guidance images. Utilizing a diffusion model as a generative prior, CODE enhances images through score-based updates along the probability-flow Ordinary Differential Equation (ODE) trajectory. This method requires no task-specific training, handcrafted modules, or assumptions, and is compatible with any diffusion model. Positioned at the intersection of conditional image generation and blind image restoration, CODE operates in a fully blind manner, relying solely on a pre-trained generative model. Our method introduces an alternative approach to blind restoration: instead of targeting a specific ground truth image based on assumptions about the underlying corruption, CODE aims to increase the likelihood of the input image while maintaining fidelity. This results in the most probable in-distribution image around the input. Our contributions are twofold. First, CODE introduces a novel editing method based on ODE providing enhanced control, realism, and fidelity compared to SDE-based counterpart. Second, we introduce a confidence interval-based clipping method, which improves CODE’s effectiveness by allowing it to disregard certain pixels or information, thus enhancing the restoration process in a blind manner. Experimental results demonstrate CODE’s effectiveness over existing methods, particularly in scenarios involving severe degradation or OoD inputs.

Predicting human trajectories is essential for the safe operation of autonomous vehicles, yet current data-driven models often lack robustness in case of noisy inputs such as adversarial examples or imperfect observations. Although some trajectory prediction methods have been developed to provide empirical robustness, these methods are heuristic and do not offer guaranteed robustness. In this work, we propose a certification approach tailored for trajectory prediction that provides guaranteed robustness. To this end, we address the unique challenges associated with trajectory prediction, such as unbounded outputs and multi-modality. To mitigate the inherent performance drop through certification, we propose a diffusion-based trajectory denoiser and integrate it into our method. Moreover, we introduce new certified performance metrics to reliably measure the trajectory prediction performance. Through comprehensive experiments, we demonstrate the accuracy and robustness of the certified predictors and highlight their advantages over the non-certified ones. The code is available online: https://s-attack.github.io/

We present GEM, a Generalizable Ego-vision Multimodal world model that predicts future frames using a reference frame, sparse features, human poses, and ego-trajectories. Hence, our model has precise control over object dynamics, ego-agent motion and human poses. GEM generates paired RGB and depth outputs for richer spatial understanding. We introduce autoregressive noise schedules to enable stable long-horizon generations. Our dataset is comprised of 4000+ hours of multimodal data across domains like autonomous driving, egocentric human activities, and drone flights. Pseudo-labels are used to get depth maps, ego-trajectories, and human poses. We use a comprehensive evaluation framework, including a new Control of Object Manipulation (COM) metric, to assess controllability. Experiments show GEM excels at generating diverse, controllable scenarios and temporal consistency over long generations. Code, models, and datasets are fully open-sourced1.

Human pose forecasting is inherently multimodal since multiple futures exist for an observed pose sequence. However, evaluating multimodality is challenging since the task is ill-posed. Therefore, we first propose an alternative paradigm to make the task well-posed. Next, while state-of-the-art methods predict multimodality, this requires oversampling a large volume of predictions. This raises key questions: (1) Can we capture multimodality by efficiently sampling fewer predictions? (2) Subsequently, which of the predicted futures is more likely for an observed pose sequence? We address these questions with MotionMap, a simple yet effective heatmap based representation for multimodality. We extend heatmaps to represent a spatial distribution over the space of all possible motions, where different local maxima correspond to different forecasts for a given observation. MotionMap can capture a variable number of modes per observation and provide confidence measures for different modes. Further, MotionMap allows us to introduce the notion of uncertainty and controllability over the forecasted pose sequence. Finally, MotionMap captures rare modes that are non-trivial to evaluate yet critical for safety. We support our claims through multiple qualitative and quantitative experiments using popular 3D human pose datasets: Human3.6M and AMASS, highlighting the strengths and limitations of our proposed method. https://vita-epfl.github.io/MotionMap

Existing vehicle trajectory prediction models struggle with generalizability, prediction uncertainties, and handling complex interactions. It is often due to limitations like complex architectures customized for a specific dataset and inefficient multimodal handling. We propose Perceiver with Register queries (PerReg+), a novel trajectory prediction framework that introduces: (1) Dual-Level Representation Learning via Self-Distillation (SD) and Masked Reconstruction (MR), capturing global context and finegrained details. Additionally, our approach of reconstructing segment-level trajectories and lane segments from masked inputs with query drop, enables effective use of contextual information and improves generalization; (2) Enhanced Multimodality using register-based queries and pretraining, eliminating the need for clustering and suppression; and (3) Adaptive Prompt Tuning during fine-tuning, freezing the main architecture and optimizing a small number of prompts for efficient adaptation. PerReg+ sets a state-of-the-art performance on nuScenes [5], Argoverse 2 [42], and Waymo Open Motion Dataset (WOMD) [13]. Remarkably, our model reduces the error by 6.8% on smaller datasets, and multi-dataset training enhances generalization. In cross-domain tests, PerReg+ reduces B-FDE by 11.8% compared to its non-pretrained variant.

Modeling spatial-temporal interactions among neighboring agents is at the heart of multi-agent problems such as motion forecasting and crowd navigation. Despite notable progress, it remains unclear to which extent modern representations can capture the causal relationships behind agent interactions. In this work, we take an in-depth look at the causal awareness of these representations, from computational formalism to real-world practice. First, we revisit the notion of non-causal robustness studied in the recent CausalAgents benchmark [56]. We show that existing representations are already partially resilient to perturbations of non-causal agents, and yet modeling indirect causal effects involving mediator agents remains challenging. To address this challenge, we introduce a metric learning approach that regularizes latent representations with causal annotations. Our controlled experiments show that this approach not only leads to higher degrees of causal awareness but also yields stronger out-of-distribution robustness. To further operationalize it in practice, we propose a sim-to-real causal transfer method via cross-domain multi-task learning. Experiments on trajectory prediction datasets show that our method can significantly boost generalization, even in the absence of real-world causal annotations, where we acquire higher prediction accuracy by only using 25% of real-world data. We hope our work provides a new perspective on the challenges and potential pathways toward causally-aware representations of multi-agent interactions. Our code is available at https://github.com/vita-epfl/CausalSim2Real.

We propose a novel fine-grained cross-view localization method that estimates the 3 Degrees of Freedom pose of a ground-level image in an aerial image of the surroundings by matching fine-grained features between the two images. The pose is estimated by aligning a point plane generated from the ground image with a point plane sampled from the aerial image. To generate the ground points, we first map ground image features to a 3D point cloud. Our method then learns to select features along the height dimension to pool the 3D points to a Bird’s-Eye-View (BEV) plane. This selection enables us to trace which feature in the ground image contributes to the BEV representation. Next, we sample a set of sparse matches from computed point correspondences between the two point planes and compute their relative pose using Procrustes alignment. Compared to the previous state-of-the-art, our method reduces the mean localization error by 28% on the VIGOR cross-area test set. Qualitative results show that our method learns semantically consistent matches across ground and aerial views through weakly supervised learning from the camera pose.

Despite progress in stereo depth estimation, omnidirectional imaging remains underexplored, mainly due to the lack of appropriate data. We introduce HELVIPAD, a real-world dataset for omnidirectional stereo depth estimation, featuring 40K video frames from video sequences across diverse environments, including crowded indoor and outdoor scenes with various lighting conditions. Collected using two 360° cameras in a top-bottom setup and a LiDAR sensor, the dataset includes accurate depth and disparity labels by projecting 3D point clouds onto equirectangular images. Additionally, we provide an augmented training set with an increased label density by using depth completion. We benchmark leading stereo depth estimation models for both standard and omnidirectional images. The results show that while recent stereo methods perform decently, a challenge persists in accurately estimating depth in omnidirectional imaging. To address this, we introduce necessary adaptations to stereo models, leading to improved performance.

Cross-view localization, the task of estimating a camera's 3-degrees-of-freedom (3-DoF) pose by aligning ground-level images with satellite images, is crucial for large-scale outdoor applications like autonomous navigation and augmented reality. Existing methods often rely on fully supervised learning, which requires costly ground-truth pose annotations. In this work, we propose GeoDistill, a Geometry guided weakly supervised self distillation framework that uses teacher-student learning with Field-of-View (FoV)-based masking to enhance local feature learning for robust cross-view localization. In GeoDistill, the teacher model localizes a panoramic image, while the student model predicts locations from a limited FoV counterpart created by FoV-based masking. By aligning the student's predictions with those of the teacher, the student focuses on key features like lane lines and ignores textureless regions, such as roads. This results in more accurate predictions and reduced uncertainty, regardless of whether the query images are panoramas or limited FoV images. Our experiments show that GeoDistill significantly improves localization performance across different frameworks. Additionally, we introduce a novel orientation estimation network that predicts relative orientation without requiring precise planar position ground truth. GeoDistill provides a scalable and efficient solution for real-world cross-view localization challenges. Code and model can be found at https://github.com/tongshw/GeoDistill.

Deep heteroscedastic regression models the mean and covariance of the target distribution through neural networks. The challenge arises from heteroscedasticity, which implies that the covariance is sample dependent and is often unknown. Consequently, recent methods learn the covariance through unsupervised frameworks, which unfortunately yield a trade-off between computational complexity and accuracy. While this trade-off could be alleviated through supervision, obtaining labels for the covariance is non-trivial. Here, we study self-supervised covariance estimation in deep heteroscedastic regression. We address two questions: (1) How should we supervise the covariance assuming ground truth is available? (2) How can we obtain pseudo labels in the absence of the ground-truth? We address (1) by analysing two popular measures: the KL Divergence and the 2-Wasserstein distance. Subsequently, we derive an upper bound on the 2-Wasserstein distance between normal distributions with non-commutative covariances that is stable to optimize. We address (2) through a simple neighborhood based heuristic algorithm which results in surprisingly effective pseudo labels for the covariance. Our experiments over a wide range of synthetic and real datasets demonstrate that the proposed 2-Wasserstein bound coupled with pseudo label annotations results in a computationally cheaper yet accurate deep heteroscedastic regression.

We propose TAROT, a targeted data selection framework grounded in Optimal Transport theory. Previous targeted data selection methods primarily rely on influence-based greedy heuristics to enhance domain-specific performance. While effective on limited, unimodal data (i.e., data following a single pattern), these methods struggle as target data complexity increases. Specifically, in multimodal distributions, such heuristics fail to account for multiple inherent patterns, leading to suboptimal data selection. This work identifies two primary limitations: (i) the disproportionate impact of dominant feature components in high-dimensional influence estimation, and (ii) the restrictive linear additive assumptions in greedy selection strategies. To address these challenges, TAROT incorporates whitened feature distance to mitigate dominant feature bias, offering a more reliable measure of data influence. Building on this, TAROT leverages whitened feature distance to quantify and minimize the optimal transport distance between selected data and target domains. Notably, this minimization also facilitates the estimation of optimal selection ratios. We evaluate TAROT across multiple tasks, including semantic segmentation, motion prediction, and instruction tuning. Results consistently show that TAROT outperforms state-of-the-art methods, demonstrating its versatility across various deep learning tasks. Code is available at: https://github.com/vita-epfl/TAROT.

Conventional human trajectory prediction models rely on clean curated data, requiring specialized equipment or manual labeling, which is often impractical for robotic applications. The existing predictors tend to overfit to clean observation affecting their robustness when used with noisy inputs. In this work, we propose MonoTransmotion (MT), a Transformerbased framework that uses only a monocular camera to jointly solve localization and prediction tasks. Our framework has two main modules: Bird's Eye View (BEV) localization and trajectory prediction. The BEV localization module estimates the position of a person using 2D human poses, enhanced by a novel directional loss for smoother sequential localizations. The trajectory prediction module predicts future motion from these estimates. We show that by jointly training both tasks with our unified framework, our method is more robust in real-world scenarios made of noisy inputs. We validate our MT network on both curated and non-curated datasets. On the curated dataset, MT achieves around 12% improvement over baseline models on BEV localization and trajectory prediction. On real-world non-curated dataset, experimental results indicate that MT maintains similar performance levels, highlighting its robustness and generalization capability. The code is available at https://github.com/vita-epfl/MonoTransmotion.

Omnidirectional depth perception is essential for mobile robotics applications that require scene understanding across a full 360° field of view. Camera-based setups offer a cost-effective option by using stereo depth estimation to generate dense, high-resolution depth maps without relying on expensive active sensing. However, existing omnidirectional stereo matching approaches achieve only limited depth accuracy across diverse environments, depth ranges, and lighting conditions, due to the scarcity of real-world data. We present DFI-OmniStereo, a novel omnidirectional stereo matching method that leverages a large-scale pre-trained foundation model for relative monocular depth estimation within an iterative optimization-based stereo matching architecture. We introduce a dedicated two-stage training strategy to utilize the relative monocular depth features for our omnidirectional stereo matching before scale-invariant fine-tuning. DFI-OmniStereo achieves state-of-the-art results on the real-world Helvipad dataset, reducing disparity MAE by approximately 16% compared to the previous best omnidirectional stereo method.

Autonomous off-road navigation faces challenges due to diverse, unstructured environments, requiring robust perception with both geometric and semantic understanding. However, scarce densely labeled semantic data limits generalization across domains. Simulated data helps, but introduces domain adaptation issues. We propose COARSE, a semi-supervised domain adaptation framework for off-road semantic segmentation, leveraging sparse, coarse in-domain labels and densely labeled out-of-domain data. Using pretrained vision transformers, we bridge domain gaps with complementary pixel-level and patch-level decoders, enhanced by a collaborative pseudo-labeling strategy on unlabeled data. Evaluations on RUGD and Rellis-3D datasets show significant improvements of 9.7% and 8.4% respectively, versus only using coarse data. Tests on real-world off-road vehicle data in a multi-biome setting further demonstrate COARSE’s applicability.

Intelligent devices for supporting persons with vision impairment are becoming more widespread, but they are lacking behind the advancements in intelligent driver assistant system. To make a first step forward, this work discusses the integration of the risk model technology, previously used in autonomous driving and advanced driver assistance systems, into an assistance device for persons with vision impairment. The risk model computes a probabilistic collision risk given object trajectories which has previously been shown to give better indications of an object's collision potential compared to distance or time-to-contact measures in vehicle scenarios. In this work, we show that the risk model is also superior in warning persons with vision impairment about dangerous objects. Our experiments demonstrate that the warning accuracy of the risk model is 67% while both distance and time-to-contact measures reach only 51% accuracy for real-world data.

We present OSKAR, the first multimodal foundation model based on bootstrapped latent feature prediction. Unlike generative or contrastive methods, it avoids memorizing unnecessary details (e.g., pixels), and does not require negative pairs, large memory banks, or hand-crafted augmentations. We propose a novel pretraining strategy: given masked tokens from multiple modalities, predict a subset of missing tokens per modality, supervised by momentum-updated uni-modal target encoders. This design efficiently utilizes the model capacity in learning high-level representations while retaining modality-specific information. Further, we propose a scalable design which decouples the compute cost from the number of modalities using a fixed representative token budget—in both input and target tokens—and introduces a parameter-efficient cross-attention predictor that grounds each prediction in the full multimodal context. We instantiate OSKAR on video, skeleton, and text modalities. Extensive experimental results show that OSKAR's unified pretrained encoder outperforms models with specialized architectures of similar size in action recognition (rgb, skeleton, frozen, low-shot) and localization, video-text retrieval, and video question answering. Project website: https://multimodal-oskar.github.io

Semantic Scene Completion (SSC) aims to reconstruct the 3D geometry and semantics of the surrounding environment. With dense voxel labels, prior works typically formulate SSC as a *dense segmentation task*, independently classifying each voxel. However, this paradigm neglects critical instance-centric discriminability, leading to instance-level incompleteness and adjacent ambiguities. To address this, we highlight a "free lunch" of SSC labels: the voxel-level class label has implicitly told the instance-level insight, which is ever-overlooked by the community. Motivated by this observation, we first introduce a training-free **Voxel-to-Instance (VoxNT) trick**: a simple yet effective method that freely converts voxel-level class labels into instance-level offset labels. Building on this, we further propose **VoxDet**, an instance-centric framework that reformulates the voxel-level SSC as *dense object detection* by decoupling it into two sub-tasks: offset regression and semantic prediction. Specifically, based on the lifted 3D volume, VoxDet first uses (a) Spatially-decoupled Voxel Encoder to generate disentangled feature volumes for the two sub-tasks, which learn task-specific spatial deformation in the densely projected tri-perceptive space. Then, we deploy (b) Task-decoupled Dense Predictor to address SSC via dense detection. Here, we first regress a 4D offset field to estimate distances (6 directions) between voxels and the corresponding object boundaries in the voxel space. The regressed offsets are then used to guide the instance-level aggregation in the classification branch, achieving instance-aware scene completion. VoxDet can be deployed on both camera and LiDAR input and jointly achieves state-of-the-art results on both benchmarks, which gives 63.0 IoU on the SemanticKITTI test set, **ranking 1$^{st}$** on the online leaderboard.

Modern language model (LM) training has been divided into multiple stages, making it difficult for downstream developers to evaluate the impact of design choices made at each stage. We present EvoLM, a model suite that enables systematic and transparent analysis of LMs' training dynamics across pre-training, continued pre-training, supervised fine-tuning, and reinforcement learning. By training over 100 LMs with 1B and 4B parameters from scratch, we rigorously evaluate both upstream (language modeling) and downstream (problem-solving) reasoning capabilities, including considerations of both in-domain and out-of-domain generalization. Key insights highlight the diminishing returns from excessive pre-training and post-training, the importance and practices of mitigating forgetting during domain-specific continued pre-training, the crucial role of continued pre-training in bridging pre-training and post-training phases, and various intricate trade-offs when configuring supervised fine-tuning and reinforcement learning. To facilitate open research and reproducibility, we release all pre-trained and post-trained models, training datasets for all stages, and our entire training and evaluation pipeline.

The generation of transferable adversarial perturbations typically involves training a generator to maximize embedding separation between clean and adversarial images at a single mid-layer of a source model. In this work, we build on this approach and introduce Neuron Attack for Transferability (NAT), a method designed to target specific neuron within the embedding. Our approach is motivated by the observation that previous layer-level optimizations often disproportionately focus on a few neurons representing similar concepts, leaving other neurons within the attacked layer minimally affected. NAT shifts the focus from embeddinglevel separation to a more fundamental, neuron-specific approach. We find that targeting individual neurons effectively disrupts the core units of the neural network, providing a common basis for transferability across different models. Through extensive experiments on 41 diverse ImageNet models and 9 fine-grained models, NAT achieves fooling rates that surpass existing baselines by over 14% in cross-model and 4% in cross-domain settings. Furthermore, by leveraging the complementary attacking capabilities of the trained generators, we achieve impressive fooling rates within just 10 queries. Our code is available at: https://krishnakanthnakka.github.io/NAT/

A major challenge to deploying robots widely is navigation in human-populated environments, commonly referred to as social robot navigation . While the field of social navigation has advanced tremendously in recent years, the fair evaluation of algorithms that tackle social navigation remains hard because it involves not just robotic agents moving in static environments but also dynamic human agents and their perceptions of the appropriateness of robot behavior. In contrast, clear, repeatable, and accessible benchmarks have accelerated progress in fields like computer vision, natural language processing and traditional robot navigation by enabling researchers to fairly compare algorithms, revealing limitations of existing solutions and illuminating promising new directions. We believe the same approach can benefit social navigation. In this article, we pave the road toward common, widely accessible, and repeatable benchmarking criteria to evaluate social robot navigation. Our contributions include (a) a definition of a socially navigating robot as one that respects the principles of safety, comfort, legibility, politeness, social competency, agent understanding, proactivity, and responsiveness to context, (b) guidelines for the use of metrics, development of scenarios, benchmarks, datasets, and simulators to evaluate social navigation, and (c) a design of a social navigation metrics framework to make it easier to compare results from different simulators, robots, and datasets.

Accurate human trajectory prediction is one of the most crucial tasks for autonomous driving, ensuring its safety. Yet, existing models often fail to fully leverage the visual cues that humans subconsciously communicate when navigating the space. In this work, we study the benefits of predicting human trajectories using human body poses instead of solely their Cartesian space locations in time. We propose ‘Social-pose’, an attention-based pose encoder that effectively captures the poses of all humans in a scene and their social relations. Our method can be integrated into various trajectory prediction architectures. We have conducted extensive experiments on state-of-the-art models (based on LSTM, GAN, MLP, and Transformer), and showed improvements over all of them on synthetic (Joint Track Auto) and real (Human3.6M, Pedestrians and Cyclists in Road Traffic, and JRDB) datasets. We also explored the advantages of using 2D versus 3D poses, as well as the effect of noisy poses and the application of our pose-based predictor in robot navigation scenarios.

Autonomous vehicles depend on accurate trajectory prediction to navigate safely in complex traffic. Yet current models are vulnerable to stealthy backdoor attacks: an adversary embeds subtle, physically plausible triggers during training that remain latent until activated. To address this risk, we introduce a structured framework categorizing four trigger types—spatial, kinetic (braking), coordinated, and composite—and demonstrate on two benchmarks (nuScenes and Argoverse 2) and two state-of-the-art architectures (Autobot and Wayformer) that poisoning as little as 5% of training samples can reliably hijack future predictions. We further propose a real-time defense leveraging social attention: by encoding agent histories, computing cross-attention to the target vehicle, and filtering out agents with anomalously high weights, our method neutralizes backdoor triggers without degrading clean-data accuracy. Comprehensive experiments show our defense reduces attack success rates across diverse urban scenarios—intersections, roundabouts, multi-lane roads—highlighting both the severity of backdoor threats and a promising pathway to secure trajectory predictors in autonomous driving systems.

Traffic forecasting is a fundamental task in transportation research, however, the scope of current research has mainly focused on a single data modality of loop detectors. Recently, the advances in Artificial Intelligence and drone technologies have enabled novel solutions to efficient, accurate, and flexible aerial observations of urban traffic. As a promising traffic monitoring approach, drones can create an accurate multi-sensor mobility observatory for large-scale urban networks, when combined with existing infrastructure. Therefore, this paper investigates urban traffic prediction from a novel perspective, where multiple input sources, i.e., loop detectors and drones, are utilized to predict both the segment-level traffic of all road segments and the regional traffic. A simple yet effective graph-based model, Hierarchical Multi-Source Neural Network (HiMSNet), is proposed to integrate multiple data modalities and learn spatio-temporal correlations. Detailed analysis shows that predicting accurate segment-level speed is more challenging than the regional speed, especially under high-demand scenarios with heavier congestion and varying traffic dynamics. Utilizing both drone and loop detector data, the prediction accuracy can be improved compared to single-modality cases, when the sensors have lower coverage and are subject to noise. Our simulation study based on vehicle trajectories in a real urban road network has highlighted the value of integrating drones in traffic forecasting and monitoring.

In contrast to vast academic efforts to study AI security, few real-world reports of AI security incidents exist. Released incidents prevent a thorough investigation of the attackers' motives, as crucial information about the company and AI application is missing. As a consequence, it often remains unknown how to avoid incidents. We tackle this gap and combine previous reports with freshly collected incidents to a small database of 32 AI security incidents. We analyze the attackers' target and goal, influencing factors, causes, and mitigations. Many incidents stem from non-compliance with best practices in security and privacy-enhancing technologies. In the case of direct AI attacks, access control may provide some mitigation, but there is little scientific work on best practices. Our paper is thus a call for action to address these gaps.

Current transformer-based skeletal action recognition models tend to focus on a limited set of joints and low-level motion patterns to predict action classes. This results in significant performance degradation under small skeleton perturbations or changing the pose estimator between training and testing. In this work, we introduce MaskCLR, a new Masked Contrastive Learning approach for Robust skeletal action recognition. We propose an Attention-Guided Proba-bilistic Masking strategy to occlude the most important joints and encourage the model to explore a larger set of discrimi-native joints. Furthermore, we propose a Multi-Level Contrastive Learning paradigm to enforce the representations of standard and occluded skeletons to be class-discriminative, i.e., more compact within each class and more dispersed across different classes. Our approach helps the model capture the high-level action semantics instead of low-level joint variations, and can be conveniently incorporated into transformer-based models. Without loss of generality, we combine MaskCLR with three transformer backbones: the vanilla transformer, DSTFormer, and STTFormer. Extensive experiments on NTU60, NTU120, and Kinetics400 show that MaskCLR consistently outperforms previous state-of-the-art methods on standard and perturbed skeletons from different pose estimators, showing improved accuracy, generalization, and robustness. Project website: https://maskclr.github.io.

Accurate human trajectory prediction is crucial for applications such as autonomous vehicles, robotics, and surveillance systems. Yet, existing models often fail to fully leverage the non-verbal social cues human subconsciously communicate when navigating the space. To address this, we introduce *Social-Transmotion*, a generic Transformer-based model that exploits diverse and numerous visual cues to predict human behavior. We translate the idea of a prompt from Natural Language Processing (NLP) to the task of human trajectory prediction, where a prompt can be a sequence of x-y coordinates on the ground, bounding boxes in the image plane, or body pose keypoints in either 2D or 3D. This, in turn, augments trajectory data, leading to enhanced human trajectory prediction. Using masking technique, our model exhibits flexibility and adaptability by capturing spatiotemporal interactions between agents based on the available visual cues. We delve into the merits of using 2D versus 3D poses, and a limited set of poses. Additionally, we investigate the spatial and temporal attention map to identify which keypoints and time-steps in the sequence are vital for optimizing human trajectory prediction. Our approach is validated on multiple datasets, including JTA, JRDB, Pedestrians and Cyclists in Road Traffic, and ETH-UCY. The code is publicly available: [https://github.com/vita-epfl/social-transmotion](https://github.com/vita-epfl/social-transmotion).

Deep heteroscedastic regression involves jointly optimizing the mean and covariance of the predicted distribution using the negative log-likelihood. However, recent works show that this may result in sub-optimal convergence due to the challenges associated with covariance estimation. While the literature addresses this by proposing alternate formulations to mitigate the impact of the predicted covariance, we focus on improving the predicted covariance itself. We study two questions: (1) Does the predicted covariance truly capture the randomness of the predicted mean? (2) In the absence of supervision, how can we quantify the accuracy of covariance estimation? We address (1) with a Taylor Induced Covariance (TIC), which captures the randomness of the predicted mean by incorporating its gradient and curvature through the second order Taylor polynomial. Furthermore, we tackle (2) by introducing a Task Agnostic Correlations (TAC) metric, which combines the notion of correlations and absolute error to evaluate the covariance. We evaluate TIC-TAC across multiple experiments spanning synthetic and real-world datasets. Our results show that not only does TIC accurately learn the covariance, it additionally facilitates an improved convergence of the negative log-likelihood. Our code is available at https://github.com/vita-epfl/TIC-TAC

Motion forecasting is crucial in enabling autonomous vehicles to anticipate the future trajectories of surrounding agents. To do so, it requires solving mapping, detection, tracking, and then forecasting problems, in a multi-step pipeline. In this complex system, advances in conventional forecasting methods have been made using curated data, i.e., with the assumption of perfect maps, detection, and tracking. This paradigm, however, ignores any errors from upstream modules. Meanwhile, an emerging end-to-end paradigm, that tightly integrates the perception and forecasting architectures into joint training, promises to solve this issue. However, the evaluation protocols between the two methods were so far incompatible and their comparison was not possible. In fact, conventional forecasting methods are usually not trained nor tested in real-world pipelines (e.g., with upstream detection, tracking, and mapping modules). In this work, we aim to bring forecasting models closer to the real-world deployment. First, we propose a unified evaluation pipeline for forecasting methods with real-world perception inputs, allowing us to compare conventional and end-to-end methods for the first time. Second, our in-depth study uncovers a substantial performance gap when transitioning from curated to perception-based data. In particular, we show that this gap (1) stems not only from differences in precision but also from the nature of imperfect inputs provided by perception modules, and that (2) is not trivially reduced by simply finetuning on perception outputs. Based on extensive experiments, we provide recommendations for critical areas that require improvement and guidance towards more robust motion forecasting in the real world. The evaluation library for benchmarking models under standardized and practical conditions is provided: https://github.com/valeoai/MFEval.

Prediction is a vital component of motion planning for autonomous vehicles (AVs). By reasoning about the possible behavior of other target agents, the ego vehicle (EV) can navigate safely, efficiently, and politely. However, most of the existing work overlooks the interdependencies of the prediction and planning module, only connecting them in a sequential pipeline or underexploring the prediction results in the planning module. In this work, we propose a framework that integrates the prediction and planning module with three highlights. First, we propose an ego-conditioned model for causal prediction, with the introduced edge-featured graph transformer model, the impact the ego future maneuver poses to the target vehicles is demonstrated. Second, we develop a motion planner based on ‘dynamic voxels’ in the spatio-temporal domain, enabling the time-to-collision criterion evaluation and the optimal trajectory generation in continuous space. Third, the prediction and planning modules are coupled in a closed-loop and efficient form. Specifically, taking each maneuver as a cluster, representative trajectory primitives are generated for conditional prediction, and conversely, prediction results are used to score the primitives as guidance, which alleviates the duplicated callback of the prediction module. The simulations are conducted in overtaking, merging, unprotected left turns, and also scenarios with imperfect social behaviors. The comparison studies demonstrate the better safety assurance and efficiency of the proposed model, and the ablation experiments further reveal the effectiveness of the new ideas.

Test-Time Adaptation (TTA) has recently gained significant attention as a new paradigm for tackling distribution shifts. Despite the sheer number of existing methods, the inconsistent experimental conditions and lack of standardization in prior literature make it difficult to measure their actual efficacies and progress. To address this issue, we present a large-scale open-sourced Test-Time Adaptation Benchmark, dubbed TTAB, which includes nine state-of-the-art algorithms, a diverse array of distribution shifts, and two comprehensive evaluation protocols. Through extensive experiments, we identify three common pitfalls in prior efforts: (i) choosing appropriate hyper-parameter, especially for model selection, is exceedingly difficult due to online batch dependency; (ii) the effectiveness of TTA varies greatly depending on the quality of the model being adapted; (iii) even under optimal algorithmic conditions, existing methods still systematically struggle with certain types of distribution shifts. Our findings suggest that future research in the field should be more transparent about their experimental conditions, ensure rigorous evaluations on a broader set of models and shifts, and re-examine the assumptions underlying the potential success of TTA for practical applications.

Predicting 3D human poses in real-world scenarios, also known as human pose forecasting, is inevitably subject to noisy inputs arising from inaccurate 3D pose estimations and occlusions. To address these challenges, we propose a diffusion-based approach that can predict given noisy observations. We frame the prediction task as a denoising problem, where both observation and prediction are considered as a single sequence containing missing elements (whether in the observation or prediction horizon). All missing elements are treated as noise and denoised with our conditional diffusion model. To better handle long-term forecasting horizon, we present a temporal cascaded diffusion model. We demonstrate the benefits of our approach on four publicly available datasets (Human3.6M, HumanEva-I, AMASS, and 3DPW), outperforming the state-of-the-art. Additionally, we show that our framework is generic enough to improve any 3D pose prediction model as a pre-processing step to repair their inputs and a post-processing step to refine their outputs. The code is available online: https://github.com/vita-epfl/DePOSit.

Scene graph generation (SGG) methods extract relationships between objects. While most methods focus on improving top-down approaches, which build a scene graph based on detected objects from an off-the-shelf object detector, there is a limited amount of work on bottom-up approaches, which jointly detect objects and their relationships in a single stage.In this work, we present a novel bottom-up SGG approach by representing relationships using Composite Relationship Fields (CoRF). CoRF turns relationship detection into a dense regression and classification task, where each cell of the output feature map identifies surrounding objects and their relationships. Furthermore, we propose a refinement head that leverages Transformers for global scene reasoning, resulting in more meaningful relationship predictions. By combining both contributions, our method outperforms previous bottom-up methods on the Visual Genome dataset by 26% while preserving real-time performance.

Occluded person re-identification (ReID) is a person retrieval task which aims at matching occluded person images with holistic ones. For addressing occluded ReID, part-based methods have been shown beneficial as they offer fine-grained information and are well suited to represent partially visible human bodies. However, training a part-based model is a challenging task for two reasons. Firstly, individual body part appearance is not as discriminative as global appearance (two distinct IDs might have the same local appearance), this means standard ReID training objectives using identity labels are not adapted to local feature learning. Secondly, ReID datasets are not provided with human topographical annotations. In this work, we propose BPBreID, a body part-based ReID model for solving the above issues. We first design two modules for predicting body part attention maps and producing body part-based features of the ReID target. We then propose GiLt, a novel training scheme for learning part-based representations that is robust to occlusions and non-discriminative local appearance. Extensive experiments on popular holistic and occluded datasets show the effectiveness of our proposed method, which outperforms state-of-the-art methods by 0.7% mAP and 5.6% rank-1 accuracy on the challenging Occluded-Duke dataset. Our code is available at https://github.com/VlSomers/bpbreid.

Human trajectory forecasting in crowds presents the challenges of modelling social interactions and outputting collision-free multimodal distribution. Following the success of Social Generative Adversarial Networks (SGAN), recent works propose various GAN-based designs to better model human motion in crowds. Despite superior performance in reducing distance-based metrics, current networks fail to output socially acceptable trajectories, as evidenced by high collisions in model predictions. To counter this, we introduce SGANv2: an improved safety-compliant SGAN architecture equipped with spatio-temporal interaction modelling and a transformer-based discriminator. The spatio-temporal modelling ability helps to learn the human social interactions better while the transformer-based discriminator design improves temporal sequence modelling. Additionally, SGANv2 utilizes the learned discriminator even at test-time via a collaborative sampling strategy that not only refines the colliding trajectories but also prevents mode collapse, a common phenomenon in GAN training. Through extensive experimentation on multiple real-world and synthetic datasets, we demonstrate the efficacy of SGANv2 to provide socially-compliant multimodal trajectories.

Vehicle trajectory prediction is nowadays a fundamental pillar of self-driving cars. Both the industry and research communities have acknowledged the need for such a pillar by providing public benchmarks. While state-of-the-art methods are impressive, i.e., they have no off-road prediction, their generalization to cities outside of the benchmark remains unexplored. In this work, we show that those methods do not generalize to new scenes. We present a method that automatically generates realistic scenes causing state-of-the-art models to go off-road. We frame the problem through the lens of adversarial scene generation. The method is a simple yet effective generative model based on atomic scene generation functions along with physical constraints. Our experiments show that more than 60% of existing scenes from the current benchmarks can be modified in a way to make prediction methods fail (i.e., predicting off-road). We further show that the generated scenes (i) are realistic since they do exist in the real world, and (ii) can be used to make existing models more robust, yielding 30-40% reductions in the off-road rate. The code is available online: https://s-attack.github.io/.

Learning behavioral patterns from observational data has been a de-facto approach to motion forecasting. Yet, the current paradigm suffers from two shortcomings: brittle under distribution shifts and inefficient for knowledge transfer. In this work, we propose to address these challenges from a causal representation perspective. We first introduce a causal formalism of motion forecasting, which casts the problem as a dynamic process with three groups of latent variables, namely invariant variables, style confounders, and spurious features. We then introduce a learning framework that treats each group separately: (i) unlike the common practice mixing datasets collected from different locations, we exploit their subtle distinctions by means of an invariance loss encouraging the model to suppress spurious correlations; (ii) we devise a modular architecture that factorizes the representations of invariant mechanisms and style confounders to approximate a sparse causal graph; (iii) we introduce a style contrastive loss that not only enforces the structure of style representations but also serves as a self-supervisory signal for test-time refinement on the fly. Experiments on synthetic and real datasets show that our proposed method improves the robustness and reusability of learned motion representations, significantly outperforming prior state-of-the-art motion forecasting models for out-of-distribution generalization and low-shot transfer.

Forecasting pedestrians' future motions is essential for autonomous driving systems to safely navigate in urban areas. However, existing prediction algorithms often overly rely on past observed trajectories and tend to fail around abrupt dynamic changes, such as when pedestrians suddenly start or stop walking. We suggest that predicting these highly non-linear transitions should form a core component to improve the robustness of motion prediction algorithms. In this paper, we introduce the new task of pedestrian stop and go forecasting. Considering the lack of suitable existing datasets for it, we release TRANS, a benchmark for explicitly studying the stop and go behaviors of pedestrians in urban traffic. We build it from several existing datasets annotated with pedestrians' walking motions, in order to have various scenarios and behaviors. We also propose a novel hybrid model that leverages pedestrian-specific and scene features from several modalities, both video sequences and high-level attributes, and gradually fuses them to integrate multiple levels of context. We evaluate our model and several baselines on TRANS, and set a new benchmark for the community to work on pedestrian stop and go forecasting.

Perceiving humans in the context of Intelligent Transportation Systems (ITS) often relies on multiple cameras or expensive LiDAR sensors. In this work, we present a new cost-effective vision-based method that perceives humans’ locations in 3D and their body orientation from a single image. We address the challenges related to the ill-posed monocular 3D tasks by proposing a neural network architecture that predicts confidence intervals in contrast to point estimates. Our neural network estimates human 3D body locations and their orientation with a measure of uncertainty. Our proposed solution (i) is privacy-safe, (ii) works with any fixed or moving cameras, and (iii) does not rely on ground plane estimation. We demonstrate the performance of our method with respect to three applications: locating humans in 3D, detecting social interactions, and verifying the compliance of recent safety measures due to the COVID-19 outbreak. We show that it is possible to rethink the concept of “social distancing” as a form of social interaction in contrast to a simple location-based rule. We publicly share the source code towards an open science mission.

Since the past few decades, human trajectory forecasting has been a field of active research owing to its numerous real-world applications: evacuation situation analysis, deployment of intelligent transport systems, traffic operations, to name a few. In this work, we cast the problem of human trajectory forecasting as learning a representation of human social interactions. Early works handcrafted this representation based on domain knowledge. However, social interactions in crowded environments are not only diverse but often subtle. Recently, deep learning methods have outperformed their handcrafted counterparts, as they learn about human-human interactions in a more generic data-driven fashion. In this work, we present an in-depth analysis of existing deep learning-based methods for modelling social interactions. We propose two domain-knowledge inspired data-driven methods to effectively capture these social interactions. To objectively compare the performance of these interaction-based forecasting models, we develop a large scale interaction-centric benchmarkTrajNet++, a significant yet missing component in the field of human trajectory forecasting. We propose novel performance metrics that evaluate the ability of a model to output socially acceptable trajectories. Experiments on TrajNet++ validate the need for our proposed metrics, and our method outperforms competitive baselines on both real-world and synthetic datasets.

Many image-based perception tasks can be formulated as detecting, associating and tracking semantic keypoints,e.g., human body pose estimation and tracking. In this work, we present a general framework that jointly detects and forms spatio-temporal keypoint associations in a single stage, making this the first real-time pose detection and tracking algorithm. We present a generic neural network architecture that uses Composite Fields to detect and construct aspatio-temporal posewhich is a single, connected graph whose nodes are the semantic keypoints (e.g., a person’s body joints) in multiple frames. For the temporal associations, we introduce the Temporal Composite Association Field (TCAF) which requires an extended network architecture and training method beyond previous Composite Fields. Our experiments show competitive accuracy while being an order of magnitude faster on multiple publicly available datasets such as COCO, CrowdPose and the PoseTrack 2017 and 2018 datasets. We also show that our method generalizes to any class of semantic keypoints such as car and animal parts to provide a holistic perception framework that is well suited for urban mobility such as self-driving cars and delivery robots.

Pedestrians are arguably one of the most safety-critical road users to consider for autonomous vehicles in urban areas. In this paper, we address the problem of jointly detecting pedestrians and recognizing 32 pedestrian attributes from a single image. These encompass visual appearance and behavior, and also include the forecasting of road crossing, which is a main safety concern. For this, we introduce a Multi-Task Learning (MTL) model relying on a composite field framework, which achieves both goals in an efficient way. Each field spatially locates pedestrian instances and aggregates attribute predictions over them. This formulation naturally leverages spatial context, making it well suited to low resolution scenarios such as autonomous driving. By increasing the number of attributes jointly learned, we highlight an issue related to the scales of gradients, which arises in MTL with numerous tasks. We solve it by normalizing the gradients coming from different objective functions when they join at the fork in the network architecture during the backward pass, referred to as fork-normalization. Experimental validation is performed on JAAD, a dataset providing numerous attributes for pedestrian analysis from autonomous vehicles, and shows competitive detection and attribute recognition results, as well as a more stable MTL training.

A powerful simulator highly decreases the need for real-world tests when training and evaluating autonomous vehicles. Data-driven simulators flourished with the recent advancement of conditional Generative Adversarial Networks (cGANs), providing high-fidelity images. The main challenge is synthesizing photorealistic images while following given constraints. In this work, we propose to improve the quality of generated images by rethinking the discriminator architecture. The focus is on the class of problems where images are generated given semantic inputs, such as scene segmentation maps or human body poses. We build on successful cGAN models to propose a new semantically-aware discriminator that better guides the generator. We aim to learn a shared latent representation that encodes enough information to jointly do semantic segmentation, content reconstruction, along with a coarse-to-fine grained adversarial reasoning. The achieved improvements are generic and simple enough to be applied to any architecture of conditional image synthesis. We demonstrate the strength of our method on the scene, building, and human synthesis tasks across three different datasets. The code is availablehttps://github.com/vita-epfl/SemDisc.

Human trajectory forecasting in crowds, at its core, is a sequence prediction problem with specific challenges of capturing inter-sequence dependencies (social interactions) and consequently predicting socially-compliant multimodal distributions. In recent years, neural network-based methods have been shown to outperform hand-crafted methods on distance-based metrics. However, these data-driven methods still suffer from one crucial limitation: lack of interpretability. To overcome this limitation, we leverage the power of discrete choice models to learn interpretable rule-based intents, and subsequently utilise the expressibility of neural networks to model scene-specific residual. Extensive experimentation on the interaction-centric benchmark TrajNet++ demonstrates the effectiveness of our proposed architecture to explain its predictions without compromising the accuracy.

Learning socially-aware motion representations is at the core of recent advances in multi-agent problems, such as human motion forecasting and robot navigation in crowds. Despite promising progress, existing representations learned with neural networks still struggle to generalize in closed-loop predictions (e.g., output colliding trajectories). This issue largely arises from the non-i.i.d. nature of sequential prediction in conjunction with ill-distributed training data. Intuitively, if the training data only comes from human behaviors in safe spaces, i.e., from "positive" examples, it is difficult for learning algorithms to capture the notion of "negative" examples like collisions. In this work, we aim to address this issue by explicitly modeling negative examples through self-supervision: (i) we intro-duce a social contrastive loss that regularizes the extracted motion representation by discerning the ground-truth positive events from synthetic negative ones; (ii) we construct informative negative samples based on our prior knowledge of rare but dangerous circumstances. Our method substantially reduces the collision rates of recent trajectory forecasting, behavioral cloning and reinforcement learning algorithms, outperforming state-of-the-art methods on several benchmarks. Our code is available at https://github.com/vita-epfl/social-nce.

Motion prediction of vehicles is critical but challenging due to the uncertainties in complex environments and the limited visibility caused by occlusions and limited sensor ranges. In this paper, we study a new task, safety-aware motion prediction with unseen vehicles for autonomous driving. Unlike the existing trajectory prediction task for seen vehicles, we aim at predicting an occupancy map that indicates the earliest time when each location can be occupied by either seen and unseen vehicles. The ability to predict unseen vehicles is critical for safety in autonomous driving. To tackle this challenging task, we propose a safety-aware deep learning model with three new loss functions to predict the earliest occupancy map. Experiments on the large-scale autonomous driving nuScenes dataset show that our proposed model significantly outperforms the state-of-the-art baselines on the safety-aware motion prediction task. To the best of our knowledge, our approach is the first one that can predict the existence of unseen vehicles in most cases. Project page at https://github.com/xrenaa/Safety-Aware-Motion-Prediction.

We present a fast bottom-up method that jointly detects over 100 keypoints on humans or objects, also referred to as human/object pose estimation. We model all keypoints belonging to a human or an object –the pose– as a graph and leverage insights from community detection to quantify the independence of keypoints. We use a graph centrality measure to assign training weights to different parts of a pose. Our proposed measure quantifies how tightly a keypoint is connected to its neighborhood. Our experiments show that our method outperforms all previous methods for human pose estimation with fine-grained keypoint annotations on the face, the hands and the feet with a total of 133 keypoints. We also show that our method generalizes to car poses.

Monocular and stereo visions are cost-effective solutions for 3D human localization in the context of self-driving cars or social robots. However, they are usually developed independently and have their respective strengths and limitations. We propose a novel unified learning framework that leverages the strengths of both monocular and stereo cues for 3D human localization. Our method jointly (i) associates humans in left- right images, (ii) deals with occluded and distant cases in stereo settings by relying on the robustness of monocular cues, and (iii) tackles the intrinsic ambiguity of monocular perspective projection by exploiting prior knowledge of the human height distribution. We specifically evaluate outliers as well as challenging instances, such as occluded and far-away pedestrians, by analyzing the entire error distribution and by estimating calibrated confidence intervals. Finally, we critically review the official KITTI 3D metrics and propose a practical 3D localization metric tailored for humans.

Test-time training (TTT) through self-supervised learning (SSL) is an emerging paradigm to tackle distributional shifts. Despite encouraging results, it remains unclear when this approach thrives or fails. In this work, we first provide an in-depth look at its limitations and show that TTT can possibly deteriorate, instead of improving, the test-time performance in the presence of severe distribution shifts. To address this issue, we introduce a test-time feature alignment strategy utilizing offline feature summarization and online moment matching, which regularizes adaptation without revisiting training data. We further scale this strategy in the online setting through batch-queue decoupling to enable robust moment estimates even with limited batch size. Given aligned feature distributions, we then shed light on the strong potential of TTT by theoretically analyzing its performance post adaptation. This analysis motivates our use of more informative self-supervision in the form of contrastive learning for visual recognition problems. We empirically demonstrate that our modified version of test-time training, termed TTT++, outperforms state-of-the-art methods by significant margins on several benchmarks. Our result indicates that storing and exploiting extra information, in addition to model parameters, can be a promising direction towards robust test-time adaptation.

The standard practice in Generative Adversarial Networks (GANs) discards the discriminator during sampling. However, this sampling method loses valuable information learned by the discriminator regarding the data distribution. In this work, we propose a collaborative sampling scheme between the generator and the discriminator for improved data generation. Guided by the discriminator, our approach refines the generated samples through gradient-based updates at a particular layer of the generator, shifting the generator distribution closer to the real data distribution. Additionally, we present a practical discriminator shaping method that can smoothen the loss landscape provided by the discriminator for effective sample refinement. Through extensive experiments on synthetic and image datasets, we demonstrate that our proposed method can improve generated samples both quantitatively and qualitatively, offering a new degree of freedom in GAN sampling.

We present an end-to-end trainable Neural Network architecture for stereo imaging that jointly locates and estimates human body poses in 3D. Our method defines a 2D pose for each human in a stereo pair of images and uses a correlation layer with a composite field to associate each left-right pair of joints. In absence of a stereo pose dataset, we show that we can train our method with synthetic data only and test it on real-world images (i.e., our training stage is domain invariant). Our method is particularly suitable for autonomous vehicles. We achieve state-of-the-art results for the 3D localization task on the challenging real-world KITTI dataset while running four times faster.

Scalable Vector Graphics (SVG) are ubiquitous in modern 2D interfaces due to their ability to scale to different resolutions. However, despite the success of deep learning-based models applied to rasterized images, the problem of vector graphics representation learning and generation remains largely unexplored. In this work, we propose a novel hierarchical generative network, called DeepSVG, for complex SVG icons generation and interpolation. Our architecture effectively disentangles high-level shapes from the low-level commands that encode the shape itself. The network directly predicts a set of shapes in a non-autoregressive fashion. We introduce the task of complex SVG icons generation by releasing a new large-scale dataset along with an open-source library for SVG manipulation. We demonstrate that our network learns to accurately reconstruct diverse vector graphics, and can serve as a powerful animation tool by performing interpolations and other latent space operations. Our code is available at https://github.com/alexandre01/deepsvg.

OBJECTIVE: Hand hygiene is essential for preventing hospital-acquired infections but is difficult to accurately track. The gold-standard (human auditors) is insufficient for assessing true overall compliance. Computer vision technology has the ability to perform more accurate appraisals. Our primary objective was to evaluate if a computer vision algorithm could accurately observe hand hygiene dispenser use in images captured by depth sensors. MATERIALS AND METHODS: Sixteen depth sensors were installed on one hospital unit. Images were collected continuously from March to August 2017. Utilizing a convolutional neural network, a machine learning algorithm was trained to detect hand hygiene dispenser use in the images. The algorithm's accuracy was then compared with simultaneous in-person observations of hand hygiene dispenser usage. Concordance rate between human observation and algorithm's assessment was calculated. Ground truth was established by blinded annotation of the entire image set. Sensitivity and specificity were calculated for both human and machine-level observation. RESULTS: A concordance rate of 96.8% was observed between human and algorithm (kappa = 0.85). Concordance among the 3 independent auditors to establish ground truth was 95.4% (Fleiss's kappa = 0.87). Sensitivity and specificity of the machine learning algorithm were 92.1% and 98.3%, respectively. Human observations showed sensitivity and specificity of 85.2% and 99.4%, respectively. CONCLUSIONS: A computer vision algorithm was equivalent to human observation in detecting hand hygiene dispenser use. Computer vision monitoring has the potential to provide a more complete appraisal of hand hygiene activity in hospitals than the current gold-standard given its ability for continuous coverage of a unit in space and time.

We present an end-to-end deep Convolutional Neural Network called Convolutional Relational Machine (CRM) for recognizing group activities that utilizes the information in spatial relations between individual persons in image or video. It learns to produce an intermediate spatial representation (activity map) based on individual and group activities. A multi-stage refinement component is responsible for decreasing the incorrect predictions in the activity map. Finally, an aggregation component uses the refined information to recognize group activities. Experimental results demonstrate the constructive contribution of the information extracted and represented in the form of the activity map. CRM shows advantages over state-of-the-art models on Volleyball and Collective Activity datasets.

We propose a new bottom-up method for multi-person 2D human pose estimation that is particularly well suited for urban mobility such as self-driving cars and delivery robots. The new method, PifPaf, uses a Part Intensity Field (PIF) to localize body parts and a Part Association Field (PAF) to associate body parts with each other to form full human poses. Our method outperforms previous methods at low resolution and in crowded, cluttered and occluded scenes thanks to (i) our new composite field PAF encoding fine-grained information and (ii) the choice of Laplace loss for regressions which incorporates a notion of uncertainty. Our architecture is based on a fully convolutional, single-shot, box-free design. We perform on par with the existing state-of-the-art bottom-up method on the standard COCO keypoint task and produce state-of-the-art results on a modified COCO keypoint task for the transportation domain.

We tackle the fundamentally ill-posed problem of 3D human localization from monocular RGB images. Driven by the limitation of neural networks outputting point estimates, we address the ambiguity in the task by predicting confidence intervals through a loss function based on the Laplace distribution. Our architecture is a light-weight feed-forward neural network that predicts 3D locations and corresponding confidence intervals given 2D human poses. The design is particularly well suited for small training data, cross-dataset generalization, and real-time applications. Our experiments show that we (i) outperform state-of-the-art results on KITTI and nuScenes datasets, (ii) even outperform a stereo-based method for far-away pedestrians, and (iii) estimate meaningful confidence intervals. We further share insights on our model of uncertainty in cases of limited observations and out-of-distribution samples.

Mobility in an effective and socially-compliant manner is an essential yet challenging task for robots operating in crowded spaces. Recent works have shown the power of deep reinforcement learning techniques to learn socially cooperative policies. However, their cooperation ability deteriorates as the crowd grows since they typically relax the problem as a one-way Human-Robot interaction problem. In this work, we want to go beyond first-order Human-Robot interaction and more explicitly model Crowd-Robot Interaction (CRI). We propose to (i) rethink pairwise interactions with a self-attention mechanism, and (ii) jointly model Human-Robot as well as Human-Human interactions in the deep reinforcement learning framework. Our model captures the Human-Human interactions occurring in dense crowds that indirectly affects the robot's anticipation capability. Our proposed attentive pooling mechanism learns the collective importance of neighboring humans with respect to their future states. Various experiments demonstrate that our model can anticipate human dynamics and navigate in crowds with time efficiency, outperforming state-of-the-art methods.

Understanding human motion behavior is critical for autonomous moving platforms (like self-driving cars and social robots) if they are to navigate human-centric environments. This is challenging because human motion is inherently multimodal: given a history of human motion paths, there are many socially plausible ways that people could move in the future. We tackle this problem by combining tools from sequence prediction and generative adversarial networks: a recurrent sequence-to-sequence model observes motion histories and predicts future behavior, using a novel pooling mechanism to aggregate information across people. We predict socially plausible futures by training adversarially against a recurrent discriminator, and encourage diverse predictions with a novel variety loss. Through experiments on several datasets we demonstrate that our approach outperforms prior work in terms of accuracy, variety, collision avoidance, and computational complexity.

We present a unified framework for understanding human social behaviors in raw image sequences. Our model jointly detects multiple individuals, infers their social actions, and estimates the collective actions with a single feed-forward pass through a neural network. We propose a single architecture that does not rely on external detection algorithms but rather is trained end-to-end to generate dense proposal maps that are refined via a novel inference scheme. The temporal consistency is handled via a person-level matching Recurrent Neural Network. The complete model takes as input a sequence of frames and outputs detections along with the estimates of individual actions and collective activities. We demonstrate state-of-the-art performance of our algorithm on multiple publicly available benchmarks.

We present an unsupervised representation learning approach that compactly encodes the motion dependencies in videos. Given a pair of images from a video clip, our framework learns to predict the long-term 3D motions. To reduce the complexity of the learning framework, we propose to describe the motion as a sequence of atomic 3D flows computed with RGB-D modality. We use a Recurrent Neural Network based Encoder-Decoder framework to predict these sequences of flows. We argue that in order for the decoder to reconstruct these sequences, the encoder must learn a robust video representation that captures long-term motion dependencies and spatial-temporal relations. We demonstrate the effectiveness of our learned temporal representations on activity classification across multiple modalities and datasets such as NTU RGB+D and MSR Daily Activity 3D. Our framework is generic to any input modality, i.e., RGB, depth, and RGB-D videos.

Physiological signals such as heart rate can provide valuable information about an individuals state and activity. However, existing work on computer vision has not yet explored leveraging these signals to enhance egocentric video understanding. In this work, we propose a model for reasoning on multimodal data to jointly predict activities and energy expenditures. We use heart rate signals as privileged self-supervision to derive energy expenditure in a training stage. A multitask objective is used to jointly optimize the two tasks. Additionally, we introduce a dataset that contains 31 hours of egocentric video augmented with heart rate and acceleration signals. This study can lead to new applications such as a visual calorie counter.

The majority of existing solutions to the Multi-Target Tracking (MTT) problem do not combine cues over a long period of time in a coherent fashion. In this paper, we present an online method that encodes long-term temporal dependencies across multiple cues. One key challenge of tracking methods is to accurately track occluded targets or those which share similar appearance properties with surrounding objects. To address this challenge, we present a structure of Recurrent Neural Networks (RNN) that jointly reasons on multiple cues over a temporal window. Our method allows to correct data association errors and recover observations from occluded states. We demonstrate the robustness of our data-driven approach by tracking multiple targets using their appearance, motion, and even interactions. Our method outperforms previous works on multiple publicly available datasets including the challenging MOT benchmark.

Existing camera networks in public spaces such as train terminals or malls can help social robots to navigate crowded scenes. However, the localization of the cameras is required, i.e., the positions and poses of all cameras in a unique reference. In this work, we estimate the relative location of any pair of cameras by solely using noisy trajectories observed from each camera. We propose a fully unsupervised learning technique using unlabelled pedestrians motion patterns captured in crowded scenes. We first estimate the pairwise camera parameters by optimally matching single-view pedestrian tracks using social awareness. Then, we show the impact of jointly estimating the network parameters. This is done by formulating a nonlinear least square optimization problem, leveraging a continuous approximation of the matching function. We evaluate our approach in real-world environments such as train terminals, where several hundreds of individuals need to be tracked across dozens of cameras every second.

Pedestrians follow different trajectories to avoid obstacles and accommodate fellow pedestrians. Any autonomous vehicle navigating such a scene should be able to foresee the future positions of pedestrians and accordingly adjust its path to avoid collisions. This problem of trajectory prediction can be viewed as a sequence generation task, where we are interested in predicting the future trajectory of people based on their past positions. Following the recent success of Recurrent Neural Network (RNN) models for sequence prediction tasks, we propose an LSTM model which can learn general human movement and predict their future trajectories. This is in contrast to traditional approaches which use hand-crafted functions such as Social forces. We demonstrate the performance of our method on several public datasets. Our model outperforms state-of-the-art methods on some of these datasets. We also analyze the trajectories predicted by our model to demonstrate the motion behaviour learned by our model.

We present an attention-based model that reasons on human body shape and motion dynamics to identify individuals in the absence of RGB information, hence in the dark. Our approach leverages unique 4D spatio-temporal signatures to address the identification problem across days. Formulated as a reinforcement learning task, our model is based on a combination of convolutional and recurrent neural networks with the goal of identifying small, discriminative regions indicative of human identity. We demonstrate that our model produces state-of-the-art results on several published datasets given only depth images. We further study the robustness of our model towards viewpoint, appearance, and volumetric changes. Finally, we share insights gleaned from interpretable 2D, 3D, and 4D visualizations of our model's spatio-temporal attention.

Inspired by the recent success of RGB-D cameras, we propose the enrichment of RGB data with an additional "quasi-free" modality, namely, the wireless signal (e.g., wifi or Bluetooth) emitted by individuals' cell phones, referred to as RGB-W. The received signal strength acts as a rough proxy for depth and a reliable cue on their identity. Although the measured signals are highly noisy (more than 2m average localization error), we demonstrate that the combination of visual and wireless data significantly improves the localization accuracy. We introduce a novel image-driven representation of wireless data which embeds all received signals onto a single image. We then indicate the ability of this additional data to (i) locate persons within a sparsity-driven framework and to (ii) track individuals with a new confidence measure on the data association problem. Our solution outperforms existing localization methods by a significant margin. It can be applied to the millions of currently installed RGB cameras to better analyze human behavior and offer the next generation of high-accuracy location-based services.

Online Multi-Object Tracking (MOT) has wide applications in time-critical video analysis scenarios, such as robot navigation and autonomous driving. In tracking-by-detection, a major challenge of online MOT is how to robustly associate noisy object detections on a new video frame with previously tracked objects. In this work, we formulate the online MOT problem as decision making in Markov Decision Processes (MDPs), where the lifetime of an object is modeled with a MDP. Learning a similarity function for data association is equivalent to learning a policy for the MDP, and the policy learning is approached in a reinforcement learning fashion which benefits from both advantages of offline-learning and online-learning for data association. Moreover, our framework can naturally handle the birth/death and appearance/disappearance of targets by treating them as state transitions in the MDP while leveraging existing online single object tracking methods. We conduct experiments on the MOT Benchmark [24] to verify the effectiveness of our method.

In crowded spaces such as city centers or train stations, human mobility looks complex, but is often influenced only by a few causes. We propose to quantitatively study crowded environments by introducing a dataset of 42 million trajectories collected in train stations. Given this dataset, we address the problem of forecasting pedestrians' destinations, a central problem in understanding large-scale crowd mobility. We need to overcome the challenges posed by a limited number of observations (e.g. sparse cameras), and change in pedestrian appearance cues across different cameras. In addition, we often have restrictions in the way pedestrians can move in a scene, encoded as priors over origin and destination (OD) preferences. We propose a new descriptor coined as Social Affinity Maps (SAM) to link broken or unobserved trajectories of individuals in the crowd, while using the OD-prior in our framework. Our experiments show improvement in performance through the use of SAM features and OD prior. To the best of our knowledge, our work is one of the first studies that provides encouraging results towards a better understanding of crowd behavior at the scale of million pedestrians.

Local Binary Descriptors are becoming more and more popular for image matching tasks, especially when going mobile. While they are extensively studied in this context, their ability to carry enough information in order to infer the original image is seldom addressed. In this work, we leverage an inverse problem approach to show that it is possible to directly reconstruct the image content from Local Binary Descriptors. This process relies on very broad assumptions besides the knowledge of the pattern of the descriptor at hand. This generalizes previous results that required either a prior learning database or non-binarized features. Furthermore, our reconstruction scheme reveals differences in the way different Local Binary Descriptors capture and encode image information. Hence, the potential applications of our work are multiple, ranging from privacy issues caused by eavesdropping image keypoints streamed by mobile devices to the design of better descriptors through the visualization and the analysis of their geometric content.

A large number of vision applications rely on matching keypoints across images. The last decade featured an arms-race towards faster and more robust keypoints and association algorithms: Scale Invariant Feature Transform (SIFT)[17], Speed-up Robust Feature (SURF)[4], and more recently Binary Robust Invariant Scalable Keypoints (BRISK)[I6] to name a few. These days, the deployment of vision algorithms on smart phones and embedded devices with low memory and computation complexity has even upped the ante: the goal is to make descriptors faster to compute, more compact while remaining robust to scale, rotation and noise. To best address the current requirements, we propose a novel keypoint descriptor inspired by the human visual system and more precisely the retina, coined Fast Retina Keypoint (FREAK). A cascade of binary strings is computed by efficiently comparing image intensities over a retinal sampling pattern. Our experiments show that FREAKs are in general faster to compute with lower memory load and also more robust than SIFT, SURF or BRISK. They are thus competitive alternatives to existing keypoints in particular for embedded applications.

Vision-based background subtraction algorithms model the intensity variation across time to classify a pixel as foreground. Unfortunately, such algorithms are sensitive to appearance changes of the background such as sudden changes of illumination or when videos are projected in the background. In this work, we propose an algorithm to extract foreground silhouettes without modeling the intensity variation across time. Using a camera pair, the stereo mismatch is processed to produce a dense disparity based on a Total Variation (TV) framework. Experimental results show that with sudden changes of background appearance, our proposed TV disparity-based extraction outperforms intensity-based algorithms and existing stereo-based approaches based on temporal depth variation and stereo mismatch.

We propose a new content-aware image resizing scheme, Stream Carving, which is based on the well-known seam carving method. Our algorithm may introduce larger seams in the retargeted image, i.e. seams with a width larger than one pixel, that we call “streams”. The resulting holes are then recovered using an inpainting method. Our retargeting algorithm is also more related to human perception by exploiting an adaptive importance map that merges several features like gradient magnitude, saliency, face, edge and straight line detection. Our approach induces an increase in the quality of the retargeted image when compared to the original seam carving method and provides similar or better results than other actual image retargeting techniques.

Typical object detection algorithms on mobile cameras suffer from the lack of a-priori knowledge on the object to be detected. The variability in the shape, pose, color distribution, and behavior affect the robustness of the detection process. In general, such variability is addressed by using a large training data. However, only objects present in the training data can be detected. This paper introduces a vision-based system to address such problem. A master-slave approach is presented where a mobile camera (the slave) can match any object detected by a fixed camera (the master). Features extracted by the master camera are used to detect the object of interest in the slave camera without the use of any training data. A single observation is enough regardless of the changes in illumination, viewpoint, color distribution and image quality. A coarse to fine description of the object is presented built upon image statistics robust to partial occlusions. Qualitative and quantitative results are presented in an indoor and an outdoor urban scene.