We introduce Human Motion Unlearning and motivate it through the concrete task of preventing violent 3D motion synthesis, an important safety requirement given that popular text-to-motion datasets (HumanML3D and Motion-X) contain from 7% to 15% violent sequences spanning both atomic gestures (e.g., a single punch) and highly compositional actions (e.g., loading and swinging a leg to kick). By focusing on violence unlearning, we demonstrate how removing a challenging, multifaceted concept can serve as a proxy for the broader capability of motion "forgetting." To enable systematic evaluation of Human Motion Unlearning, we establish the first motion unlearning benchmark by automatically filtering HumanML3D and Motion-X datasets to create distinct forget sets (violent motions) and retain sets (safe motions). We introduce evaluation metrics tailored to sequential unlearning, measuring both suppression efficacy and the preservation of realism and smooth transitions. We adapt two state-of-the-art, training-free image unlearning methods (UCE and RECE) to leading text-to-motion architectures (MoMask and BAMM), and propose Latent Code Replacement (LCR), a novel, training-free approach that identifies violent codes in a discrete codebook representation and substitutes them with safe alternatives. Our experiments show that unlearning violent motions is indeed feasible and that acting on latent codes strikes the best trade-off between violence suppression and preserving overall motion quality. This work establishes a foundation for advancing safe motion synthesis across diverse applications.

Intention drives human movement in complex environments, but such movement can only happen if the surrounding context supports it. Despite the intuitive nature of this mechanism, existing research has not yet provided tools to evaluate the alignment between skeletal movement (motion), intention (text), and the surrounding context (scene). In this work, we introduce MonSTeR, the first MOtioN-Scene-TExt Retrieval model. Inspired by the modeling of higher-order relations, MonSTeR constructs a unified latent space by leveraging unimodal and cross-modal representations. This allows MonSTeR to capture the intricate dependencies between modalities, enabling flexible but robust retrieval across various tasks. Our results show that MonSTeR outperforms trimodal models that rely solely on unimodal representations. Furthermore, we validate the alignment of our retrieval scores with human preferences through a dedicated user study. We demonstrate the versatility of MonSTeR's latent space on zero-shot in-Scene Object Placement and Motion Captioning. Code and pre-trained models are available at github.com/colloroneluca/MonSTeR.

Image-text representation learning forms a cornerstone in vision-language models, where pairs of images and textual descriptions are contrastively aligned in a shared embedding space. Since visual and textual concepts are naturally hierarchical, recent work has shown that hyperbolic space can serve as a high-potential manifold to learn vision-language representation with strong downstream performance. In this work, for the first time we show how to fully leverage the innate hierarchical nature of hyperbolic embeddings by looking beyond individual image-text pairs. We propose Compositional Entailment Learning for hyperbolic vision-language models. The idea is that an image is not only described by a sentence but is itself a composition of multiple object boxes, each with their own textual description. Such information can be obtained freely by extracting nouns from sentences and using openly available localized grounding models. We show how to hierarchically organize images, image boxes, and their textual descriptions through contrastive and entailment-based objectives. Empirical evaluation on a hyperbolic vision-language model trained with millions of image-text pairs shows that the proposed compositional learning approach outperforms conventional Euclidean CLIP learning, as well as recent hyperbolic alternatives, with better zero-shot and retrieval generalization and clearly stronger hierarchical performance.

The success of collaboration between humans and robots in shared environments relies on the robot's real-time adaptation to human motion. Specifically, in Social Navigation, the agent should be close enough to assist but ready to back up to let the human move freely, avoiding collisions. Human trajectories emerge as crucial cues in Social Navigation, but they are partially observable from the robot's egocentric view and computationally complex to process. We present the first Social Dynamics Adaptation model (SDA) based on the robot's state-action history to infer the social dynamics. We propose a two-stage Reinforcement Learning framework: the first learns to encode the human trajectories into social dynamics and learns a motion policy conditioned on this encoded information, the current status, and the previous action. Here, the trajectories are fully visible, i.e., assumed as privileged information. In the second stage, the trained policy operates without direct access to trajectories. Instead, the model infers the social dynamics solely from the history of previous actions and statuses in real-time. Tested on the novel Habitat 3.0 platform, SDA sets a novel state-of-the-art (SotA) performance in finding and following humans. The code can be found at https://github.com/L-Scofano/SDA.

We introduce the novel task of Crowd Volume Estimation (C VE), defined as the process of estimating the collective body volume of crowds using only RGB images. Besides event management and public safety, CVE can be instrumental in approximating body weight, unlocking weight-sensitive applications such as infrastructure stress assessment, and assuring even weight balance. We propose the first benchmark for CVE, comprising ANTHROPOS-V, a synthetic photorealistic video dataset featuring crowds in diverse urban environments. Its annotations include each person's volume, SMPL shape parameters, and key-points. Also, we explore metrics pertinent to CVE, de-fine baseline models adapted from Human Mesh Recov-ery and Crowd Counting domains, and propose a CVE-specific methodology that surpasses baselines. Although synthetic, the weights and heights of individuals are aligned with the real-world population distribution across gen-ders, and they transfer to the downstream task of CVE from real images. Benchmark and code are available at github.com/colloronelucaICrowd- Volume-Estimation.

Accurately estimating the 3D pose of the camera wearer in egocentric video sequences is crucial to modeling human behavior in virtual and augmented reality applications. The task presents unique challenges due to the limited visibility of the user's body caused by the front-facing camera mounted on their head. Recent research has explored the utilization of the scene and ego-motion, but it has overlooked humans' interactive nature. We propose a novel framework for Social Egocentric Estimation of body MEshes (SEE-ME). Our approach is the first to estimate the wearer's mesh using only a latent probabilistic diffusion model, which we condition on the scene and, for the first time, on the social wearer-interactee interactions. Our in-depth study sheds light on when social interaction matters most for ego-mesh estimation; it quantifies the impact of interpersonal distance and gaze direction. Overall, SEEME surpasses the current best technique, reducing the pose estimation error (MPJPE) by 53%. The code is available at SEEME.

Promptly identifying procedural errors from egocentric videos in an online setting is highly challenging and valuable for detecting mistakes as soon as they happen. This capability has a wide range of applications across various fields, such as manufacturing and healthcare. The nature of procedural mistakes is open-set since novel types of failures might occur, which calls for one-class classifiers trained on correctly executed procedures. However, no technique can currently detect open-set procedural mistakes online. We propose PREGO, the first online one-class classification model for mistake detection in PRocedural EGOcentric videos. PREGO is based on an online action recognition component to model the current action, and a symbolic reasoning module to predict the next actions. Mistake detection is performed by comparing the recognized current action with the expected future one. We evaluate PREGO on two procedural egocentric video datasets, Assembly101 and Epic-tent, which we adapt for online benchmarking of procedural mistake detection to establish suitable benchmarks, thus defining the Assembly101-O and Epic-tent-O datasets, respectively. The code is available at https://github.com/alef/abo/PREGO.

We introduce a hyperbolic neural network approach to pixel-level active learning for semantic segmentation. Analysis of the data statistics leads to a novel interpretation of the hyperbolic radius as an indicator of data scarcity. In HALO (Hyperbolic Active Learning Optimization), for the first time, we propose the use of epistemic uncertainty as a data acquisition strategy, following the intuition of selecting data points that are the least known. The hyperbolic radius, complemented by the widely-adopted prediction entropy, effectively approximates epistemic uncertainty. We perform extensive experimental analysis based on two established synthetic-to-real benchmarks, i.e. GTAV $\rightarrow$ Cityscapes and SYNTHIA $\rightarrow$ Cityscapes. Additionally, we test HALO on Cityscape $\rightarrow$ ACDC for domain adaptation under adverse weather conditions, and we benchmark both convolutional and attention-based backbones. HALO sets a new state-of-the-art in active learning for semantic segmentation under domain shift and it is the first active learning approach that surpasses the performance of supervised domain adaptation while using only a small portion of labels (i.e., 1%).

Autonomous robots are increasingly becoming a strong fixture in social environments. Effective crowd navigation requires not only safe yet fast planning, but should also enable interpretability and computational efficiency for working in real-time on embedded devices. In this work, we advocate for hyperbolic learning to enable crowd navigation and we introduce Hyp2Nav. Different from conventional reinforcement learning-based crowd navigation methods, Hyp2Nav leverages the intrinsic properties of hyperbolic geometry to better encode the hierarchical nature of decision-making processes in navigation tasks. We propose a hyperbolic policy model and a hyperbolic curiosity module that results in effective social navigation, best success rates, and returns across multiple simulation settings, using up to 6 times fewer parameters than competitor state-of-the-art models. With our approach, it becomes even possible to obtain policies that work in 2-dimensional embedding spaces, opening up new possibilities for low-resource crowd navigation and model interpretability. Insightfully, the internal hyperbolic representation of Hyp2Nav correlates with how much attention the robot pays to the surrounding crowds, e.g. due to multiple people occluding its pathway or to a few of them showing colliding plans, rather than to its own planned route. The code is available at https://github.com/GDam90/hyp2nav.

Abstract Forecasting players in sports has grown in popularity due to the potential for a tactical advantage and the applicability of such research to multi-agent interaction systems. Team sports contain a significant social component that influences interactions between teammates and opponents. However, it still needs to be fully exploited. In this work, we hypothesize that each participant has a specific function in each action and that role-based interaction is critical for predicting players’ future moves. We create RolFor, a novel end-to-end model for Role-based Forecasting. RolFor uses a new module we developed called Ordering Neural Networks (OrderNN) to permute the order of the players such that each player is assigned to a latent role. The latent role is then modeled with a RoleGCN. Thanks to its graph representation, it provides a fully learnable adjacency matrix that captures the relationships between roles and is subsequently used to forecast the players’ future trajectories. Extensive experiments on a challenging NBA basketball dataset back up the importance of roles and justify our goal of modeling them using optimizable models. When an oracle provides roles, the proposed RolFor compares favorably to the current state-of-the-art (it ranks first in terms of ADE and second in terms of FDE errors). However, training the end-to-end RolFor incurs the issues of differentiability of permutation methods, which we experimentally review. Finally, this work restates differentiable ranking as a difficult open problem and its great potential in conjunction with graph-based interaction models.

Detecting the anomaly of human behavior is paramount to timely recognizing endangering situations, such as street fights or elderly falls. However, anomaly detection is complex since anomalous events are rare and because it is an open set recognition task, i.e., what is anomalous at inference has not been observed at training. We propose COSKAD, a novel model that encodes skeletal human motion by a graph convolutional network and learns to COntract SKeletal kinematic embeddings onto a latent hypersphere of minimum volume for Video Anomaly Detection. We propose three latent spaces: the commonly-adopted Euclidean and the novel spherical and hyperbolic. All variants outperform the state-of-the-art on the most recent UBnormal dataset, for which we contribute a human-related version with annotated skeletons. COSKAD sets a new state-of-the-art on the human-related versions of ShanghaiTech Campus and CUHK Avenue, with performance comparable to video-based methods. Source code and dataset will be released upon acceptance.

This paper proposes the combination of a fast motion planner, which generates potentially unfeasible trajectories for the center of mass of an underactuated UAV, with a tracking controller integrating a Control Barrier Function. In this way avoidance of both stationary and moving obstacles is guaranteed even when the UAV deviates from the collision-free trajectory issued by the planner, due to its unfeasibility. Simulations in a dynamic and cluttered environment validate the method.

Anomalies are rare and anomaly detection is often therefore framed as One-Class Classification (OCC), i.e. trained solely on normalcy. Leading OCC techniques constrain the latent representations of normal1motions to limited volumes and detect as abnormal anything outside, which accounts satisfactorily for the openset’ness of anomalies. But normalcy shares the same openset’ness property since humans can perform the same action in several ways, which the leading techniques neglect.We propose a novel generative model for video anomaly detection (VAD), which assumes that both normality and abnormality are multimodal. We consider skeletal representations and leverage state-of-the-art diffusion probabilistic models to generate multimodal future human poses. We contribute a novel conditioning on the past motion of people and exploit the improved mode coverage capabilities of diffusion processes to generate different-but-plausible future motions. Upon the statistical aggregation of future modes, an anomaly is detected when the generated set of motions is not pertinent to the actual future. We validate our model on 4 established benchmarks: UBnormal, HR-UBnormal, HR-STC, and HR-Avenue, with extensive experiments surpassing state-of-the-art results.

Deep-learning models for 3D point cloud semantic segmentation exhibit limited generalization capabilities when trained and tested on data captured with different sensors or in varying environments due to domain shift. Domain adaptation methods can be employed to mitigate this domain shift, for instance, by simulating sensor noise, developing domain-agnostic generators, or training point cloud completion networks. Often, these methods are tailored for range view maps or necessitate multi-modal input. In contrast, domain adaptation in the image domain can be executed through sample mixing, which emphasizes input data manipulation rather than employing distinct adaptation modules. In this study, we introduce compositional semantic mixing for point cloud domain adaptation, representing the first unsupervised domain adaptation technique for point cloud segmentation based on semantic and geometric sample mixing. We present a two-branch symmetric network architecture capable of concurrently processing point clouds from a source domain (e.g. synthetic) and point clouds from a target domain (e.g. real-world). Each branch operates within one domain by integrating selected data fragments from the other domain and utilizing semantic information derived from source labels and target (pseudo) labels. Additionally, our method can leverage a limited number of human point-level annotations (semi-supervised) to further enhance performance. We assess our approach in both synthetic-to-real and real-to-real scenarios using LiDAR datasets and demonstrate that it significantly outperforms state-of-the-art methods in both unsupervised and semi-supervised settings.

Transformer Networks have established themselves as the de-facto state-of-the-art for trajectory forecasting but there is currently no systematic study on their capability to model the motion patterns of people, without interactions with other individuals nor the social context. There is abundant literature on LSTMs, CNNs and GANs on this subject. However methods adopting Transformer techniques achieve great performances by complex models and a clear analysis of their adoption as plain sequence models is missing. This paper proposes the first in-depth study of Transformer Networks (TF) and the Bidirectional Transformers (BERT) for the forecasting of the individual motion of people, without bells and whistles. We conduct an exhaustive evaluation of the input/output representations, problem formulations and sequence modelling, including a novel analysis of their capability to predict multi-modal futures. Out of comparative evaluation on the ETH+UCY benchmark, both TF and BERT are top performers in predicting individual motions and remain within a narrow margin wrt more complex techniques, including both social interactions and scene contexts. Source code will be released for all conducted experiments.

Unsupervised Domain Adaptation (UDA) is a key issue in visual recognition, as it allows to bridge different visual domains enabling robust performances in the real world. To date, all proposed approaches rely on human expertise to manually adapt a given UDA method (e.g. DANN) to a specific backbone architecture (e.g. ResNet). This dependency on handcrafted designs limits the applicability of a given approach in time, as old methods need to be constantly adapted to novel backbones.Existing Neural Architecture Search (NAS) approaches cannot be directly applied to mitigate this issue, as they rely on labels that are not available in the UDA setting. Furthermore, most NAS methods search for full architectures, which precludes the use of pre-trained models, essential in a vast range of UDA settings for reaching SOTA results.To the best of our knowledge, no prior work has addressed these aspects in the context of NAS for UDA. Here we tackle both aspects with an Adversarial Branch Architecture Search for UDA (ABAS): i. we address the lack of target labels by a novel data-driven ensemble approach for model selection; and ii. we search for an auxiliary adversarial branch, attached to a pre-trained backbone, which drives the domain alignment. We extensively validate ABAS to improve two modern UDA techniques, DANN and ALDA, on three standard visual recognition datasets (Office31, Office-Home and PACS). In all cases, ABAS robustly finds the adversarial branch architectures and parameters which yield best performances. https://github.com/lr94/abas.

Human pose forecasting is a complex structured-data sequence-modelling task, which has received increasing attention, also due to numerous potential applications. Research has mainly addressed the temporal dimension as time series and the interaction of human body joints with a kinematic tree or by a graph. This has decoupled the two aspects and leveraged progress from the relevant fields, but it has also limited the understanding of the complex structural joint spatio-temporal dynamics of the human pose.Here we propose a novel Space-Time-Separable Graph Convolutional Network (STS-GCN) for pose forecasting. For the first time, STS-GCN models the human pose dynamics only with a graph convolutional network (GCN), including the temporal evolution and the spatial joint interaction within a single-graph framework, which allows the cross-talk of motion and spatial correlations. Concurrently, STS-GCN is the first space-time-separable GCN: the space-time graph connectivity is factored into space and time affinity matrices, which bottlenecks the space-time cross-talk, while enabling full joint-joint and time-time correlations. Both affinity matrices are learnt end-to-end, which results in connections substantially deviating from the standard kinematic tree and the linear-time time series.In experimental evaluation on three complex, recent and large-scale benchmarks, Human3.6M [24], AMASS [34] and 3DPW [48], STS-GCN outperforms the state-of-the-art, surpassing the current best technique [35] by over 32% in average at the most difficult long-term predictions, while only requiring 1.7% of its parameters. We explain the results qualitatively and illustrate the graph interactions by the factored joint-joint and time-time learnt graph connections. Our source code is available at: https://github.com/FraLuca/STSGCN

In this article, we explore the correlation between people trajectories and their head orientations. We argue that people trajectory and head pose forecasting can be modelled as a joint problem. Recent approaches on trajectory forecasting leverage short-term trajectories (aka tracklets) of pedestrians to predict their future paths. In addition, sociological cues, such as expected destination or pedestrian interaction, are often combined with tracklets. In this article, we propose MiXing-LSTM (MX-LSTM) to capture the interplay between positions and head orientations (vislets) thanks to a joint unconstrained optimization of full covariance matrices during the LSTM backpropagation. We additionally exploit the head orientations as a proxy for the visual attention, when modeling social interactions. MX-LSTM predicts future pedestrians location and head pose, increasing the standard capabilities of the current approaches on long-term trajectory forecasting. Compared to the state-of-the-art, our approach shows better performances on an extensive set of public benchmarks. MX-LSTM is particularly effective when people move slowly, i.e., the most challenging scenario for all other models. The proposed approach also allows for accurate predictions on a longer time horizon.

3D object detectors based only on LiDAR point clouds hold the state-of-the-art on modern street-view benchmarks. However, LiDAR-based detectors poorly generalize across domains due to domain shift. In the case of LiDAR, in fact, domain shift is not only due to changes in the environment and in the object appearances, as for visual data from RGB cameras, but is also related to the geometry of the point clouds (e.g., point density variations). This paper proposes SF-UDA3D, the first Source-Free Unsupervised Domain Adaptation (SF-UDA) framework to domain-adapt the state-of-the-art PointRCNN 3D detector to target domains for which we have no annotations (unsupervised), neither we hold images nor annotations of the source domain (source-free). SF-UDA3Dis novel on both aspects. Our approach is based on pseudo-annotations, reversible scale-transformations and motion coherency. SFUDA3Doutperforms both previous domain adaptation techniques based on features alignment and state-of-the-art 3D object detection methods which additionally use few-shot target annotations or target annotation statistics. This is demonstrated by extensive experiments on two large-scale datasets, i.e., KITTI and nuScenes.

Most recent successes on forecasting the people motion are based on LSTM models and all most recent progress has been achieved by modelling the social interaction among people and the people interaction with the scene. We question the use of the LSTM models and propose the novel use of Transformer Networks for trajectory forecasting. This is a fundamental switch from the sequential step-by-step processing of LSTMs to the only-attention-based memory mechanisms of Transformers. In particular, we consider both the original Transformer Network (TF) and the larger Bidirectional Transformer (BERT), state-of-the-art on all natural language processing tasks. Our proposed Transformers predict the trajectories of the individual people in the scene. These are “simple” models because each person is modelled separately without any complex human-human nor scene interaction terms. In particular, the TF modelwithoutbellsandwhistlesyields the best score on the largest and most challenging trajectory forecasting benchmark of TrajNet [1]. Additionally, its extension which predicts multiple plausible future trajectories performs on par with more engineered techniques on the 5 datasets of ETH [2]+UCY [3]. Finally, we show that Transformers may deal with missing observations, as it may be the case with real sensor data. Code is available at github.com/FGiuliari/Trajectory-Transformer.

Person search has recently gained attention as the novel task of finding a person, provided as a cropped sample, from a gallery of non-cropped images, whereby several other people are also visible. We believe that i. person detection and re-identification should be pursued in a joint optimization framework and that ii. the person search should leverage the query image extensively (e.g. emphasizing unique query patterns). However, so far, no prior art realizes this. We introduce a novel query-guided end-to-end person search network (QEEPS) to address both aspects. We leverage a most recent joint detector and re-identification work, OIM [37]. We extend this with i. a query-guided Siamese squeeze-and-excitation network (QSSE-Net) that uses global context from both the query and gallery images, ii. a query-guided region proposal network (QRPN) to produce query-relevant proposals, and iii. a query-guided similarity subnetwork (QSimNet), to learn a query-guided re-identification score. QEEPS is the first end-to-end query-guided detection and re-id network. On both the most recent CUHK-SYSU [37] and PRW [46] datasets, we outperform the previous state-of-the-art by a large margin.

Lighting design in indoor environments is of primary importance for at least two reasons: 1) people should perceive an adequate light; 2) an effective lighting design means consistent energy saving. We present the Invisible Light Switch (ILS) to address both aspects. ILS dynamically adjusts the room illumination level to save energy while maintaining constant the light level perception of the users. So the energy saving is invisible to them. Our proposed ILS leverages a radiosity model to estimate the light level which is perceived by a person within an indoor environment, taking into account the person position and her/his viewing frustum (head pose). ILS may therefore dim those luminaires, which are not seen by the user, resulting in an effective energy saving, especially in large open offices (where light may otherwise be ON everywhere for a single person). To quantify the system performance, we have collected a new dataset where people wear luxmeter devices while working in office rooms. The luxmeters measure the amount of light (in Lux) reaching the people gaze, which we consider a proxy to their illumination level perception. Our initial results are promising: in a room with 8 LED luminaires, the energy consumption in a day may be reduced from 18585 to 6206 watts with ILS (currently needing 1560 watts for operations). While doing so, the drop in perceived lighting decreases by just 200 lux, a value considered negligible when the original illumination level is above 1200 lux, as is normally the case in offices.

Lighting design and modelling or industrial applications like luminaire planning and commissioning rely heavily on time-consuming manual measurements or on physically coherent computational simulations. Regarding the latter, standard approaches are based on CAD modeling simulations and offline rendering, with long processing times and therefore inflexible workflows. Thus, in this paper we propose a computer vision based system to measure lighting with just a single RGBD camera. The proposed method uses both depth data and images from the sensor to provide a dense measure of light intensity in the field of view of the camera. We evaluate our system on novel ground truth data and compare it to state-of-the-art commercial light planning software. Our system provides improved performance, while being completely automated, given that the CAD model is extracted from the depth and the albedo estimated with the support of RGB images. To the best of our knowledge, this is the first automatic framework for the estimation of lighting in general indoor scenarios from RGBD input.

A desirable smart traffic-monitoring and street-safety system can elicit and support the intervention of law enforcement agencies or medical staff. Recently, there has been a dramatically higher demand for such smart systems. To this end, the International Workshop on Traffic and Street Surveillance for Safety and Security (IWT4S) was organized in conjunction with the 15th IEEE International Conference on Advanced Video and Signal-based Surveillance (AVSS 2018). Our goal is to advance the state-of-the-art detection and tracking algorithms and provide a comprehensive performance evaluation for them. We evaluate 5 submitted detection and 7 submitted tracking methods on the large-scale UA-DETRAC benchmark, and the results are shared publicly on the website http://detrac-db. rit.albany.edu. We expect this challenge to advance the research and development of new detection and tracking methods for transportation applications.

Recent approaches on trajectory forecasting use tracklets to predict the future positions of pedestrians exploiting Long Short Term Memory (LSTM) architectures. This paper shows that adding vislets, that is, short sequences of head pose estimations, allows to increase significantly the trajectory forecasting performance. We then propose to use vislets in a novel framework called MX-LSTM, capturing the interplay between tracklets and vislets thanks to a joint unconstrained optimization of full covariance matrices during the LSTM backpropagation. At the same time, MX-LSTM predicts the future head poses, increasing the standard capabilities of the long-term trajectory forecasting approaches. With standard head pose estimators and an attentional-based social pooling, MX-LSTM scores the new trajectory forecasting state-of-the-art in all the considered datasets (Zara01, Zara02, UCY, and TownCentre) with a dramatic margin when the pedestrians slow down, a case where most of the forecasting approaches struggle to provide an accurate solution.

In this paper we show the importance of the head pose estimation in the task of trajectory forecasting. This cue, when produced by an oracle and injected in a novel socially-based energy minimization approach, allows to get state-of-the-art performances on four different forecasting benchmarks, without relying on additional information such as expected destination and desired speed, which are supposed to be know beforehand for most of the current forecasting techniques. Our approach uses the head pose estimation for two aims: 1) to define a view frustum of attention, highlighting the people a given subject is more interested about, in order to avoid collisions; 2) to give a shorttime estimation of what would be the desired destination point. Moreover, we show that when the head pose estimation is given by a real detector, though the performance decreases, it still remains at the level of the top score forecasting systems.

Faster R-CNN has established itself as the de-facto best object detector but it remains strongly limited in two aspects: (i) it is sensitive to background clutter and its classification performance decreases when it is confronted with more noisy proposals; (ii) it suffers when the objects vary largely in scale and specifically for the small objects. We address both issues with our geometric-proposals for Faster R-CNN (GP-FRCNN), whereby we re-rank the generic object proposals with an approximate geometric estimate of the scene. But the devil is in the details: the simple extension requires involved scale adjustments (e.g. anchors, layer resolution) which we detail in this paper. Finally, our GP-FRCNN performs equally well on smaller and larger objects, a long standing challenge for any object detection algorithm. The application of GP-FRCNN to surveillance videos is straightforward and does not require an explicit geometric formulation. We extensively test the model on the UA-DETRAC dataset, where GP-FRCNN outperforms the standard Faster R-CNN by 19%.

The rapid advances of transportation infrastructure have led to a dramatic increase in the demand for smart systems capable of monitoring traffic and street safety. Fundamental to these applications are a community-based evaluation platform and benchmark for object detection and multi-object tracking. To this end, we organize the AVSS2017 Challenge on Advanced Traffic Monitoring, in conjunction with the International Workshop on Traffic and Street Surveillance for Safety and Security (IWT4S), to evaluate the state-of-the-art object detection and multi-object tracking algorithms in the relevance of traffic surveillance. Submitted algorithms are evaluated using the large-scale UA-DETRAC benchmark and evaluation protocol. The benchmark, the evaluation toolkit and the algorithm performance are publicly available from the website http://detrac-db.rit.albany.edu.

The head pose is an important cue for computer vision. Traditionally considered in human computer interaction applications, it becomes very hard to model in surveillance scenarios, due to the tiny head size. Additionally, no public dataset contains continuous head pose annotations in open scenery, making the challenge even harder to face. Here we present a framework based on Faster RCNN, which introduces a branch in the network architecture related to the head pose estimation. The key idea is to leverage the presence of the people body to better infer the head pose, through a joint optimization process. Additionally, we enrich the Town Center dataset with head pose labels, promoting further study on this topic. Results on this novel benchmark and ablation studies on other task-specific datasets promote our idea and confirm the importance of the body cues to contextualize the head pose estimation.

Video segmentation has become an important and active research area with a large diversity of proposed approaches. Graph-based methods, enabling top-performance on recent benchmarks, consist of three essential components: 1. powerful features account for object appearance and motion similarities; 2. spatio-temporal neighborhoods of pixels or superpixels (the graph edges) are modeled using a combination of those features; 3. video segmentation is formulated as a graph partitioning problem. While a wide variety of features have been explored and various graph partition algorithms have been proposed, there is surprisingly little research on how to construct a graph to obtain the best video segmentation performance. This is the focus of our paper. We propose to combine features by means of a classifier, use calibrated classifier outputs as edge weights and define the graph topology by edge selection. By learning the graph (without changes to the graph partitioning method), we improve the results of the best performing video segmentation algorithm by 6% on the challenging VSB100 benchmark, while reducing its runtime by 55%, as the learnt graph is much sparser.

Computational and memory costs restrict spectral techniques to rather small graphs, which is a serious limitation especially in video segmentation. In this paper, we propose the use of a reduced graph based on superpixels. In contrast to previous work, the reduced graph is reweighted such that the resulting segmentation is equivalent, under certain assumptions, to that of the full graph. We consider equivalence in terms of the normalized cut and of its spectral clustering relaxation. The proposed method reduces runtime and memory consumption and yields on par results in image and video segmentation. Further, it enables an efficient data representation and update for a new streaming video segmentation approach that also achieves state-of-the-art performance.

Video segmentation research is currently limited by the lack of a benchmark dataset that covers the large variety of sub problems appearing in video segmentation and that is large enough to avoid over fitting. Consequently, there is little analysis of video segmentation which generalizes across subtasks, and it is not yet clear which and how video segmentation should leverage the information from the still-frames, as previously studied in image segmentation, alongside video specific information, such as temporal volume, motion and occlusion. In this work we provide such an analysis based on annotations of a large video dataset, where each video is manually segmented by multiple persons. Moreover, we introduce a new volume-based metric that includes the important aspect of temporal consistency, that can deal with segmentation hierarchies, and that reflects the tradeoff between over-segmentation and segmentation accuracy.

We propose a novel model for the spatio-temporal clustering of trajectories based on motion, which applies to challenging street-view video sequences of pedestrians captured by a mobile camera. A key contribution of our work is the introduction of novel probabilistic region trajectories, motivated by the non-repeatability of segmentation of frames in a video sequence. Hierarchical image segments are obtained by using a state-of-the-art hierarchical segmentation algorithm, and connected from adjacent frames in a directed acyclic graph. The region trajectories and measures of confidence are extracted from this graph using a dynamic programming-based optimisation. Our second main contribution is a Bayesian framework with a twofold goal: to learn the optimal, in a maximum likelihood sense, Random Forests classifier of motion patterns based on video features, and construct a unique graph from region trajectories of different frames, lengths and hierarchical levels. Finally, we demonstrate the use of Isomap for effective spatio-temporal clustering of the region trajectories of pedestrians. We support our claims with experimental results on new and existing challenging video sequences.

This paper proposes a probabilistic graphical model for the problem of propagating labels in video sequences, also termed the label propagation problem. Given a limited amount of hand labelled pixels, typically the start and end frames of a chunk of video, an EM based algorithm propagates labels through the rest of the frames of the video sequence. As a result, the user obtains pixelwise labelled video sequences along with the class probabilities at each pixel. Our novel algorithm provides an essential tool to reduce tedious hand labelling of video sequences, thus producing copious amounts of useable ground truth data. A novel application of this algorithm is in semi-supervised learning of discriminative classifiers for video segmentation and scene parsing. The label propagation scheme can be based on pixel-wise correspondences obtained from motion estimation, image patch based similarities as seen in epitomic models or even the more recent, semantically consistent hierarchical regions. We compare the abilities of each of these variants, both via quantitative and qualitative studies against ground truth data. We then report studies on a state of the art Random forest classifier based video segmentation scheme, trained using fully ground truth data and with data obtained from label propagation. The results of this study strongly support and encourage the use of the proposed label propagation algorithm.

Since the pioneering work of Gibson in 1950, Shape-From-Texture has been considered by researchers as a hard problem, mainly due to restrictive assumptions which often limit its applicability. We assume a very general stochastic homogeneity and perspective camera model, for both deterministic and stochastic textures. A multi-scale distortion is efficiently estimated with a previously presented method based on Fourier analysis and Gabor filters. The novel 3D reconstruction method that we propose applies to general shapes, and includes non-developable and extensive surfaces. Our algorithm is accurate, robust and compares favorably to the present state of the art of Shape-From-Texture. Results show its application to non-invasively study shape changes with laid-on textures, while rendering and re-texturing of cloth is suggested for future work.

Shape from texture has received much attention in the past few decades. We propose a computationally efficient method to extract 3D planar surface orientation from the spectral variations of a visual texture. Under the assumption of homogeneity, the texture is represented by the novel method of identifying ridges of its Fourier transform. Local spatial frequencies are then computed using a minimal set of selected Gabor filters. Under perspective projection, frequencies are backprojected and orientation is computed so as to minimize the variance of the frequencies' backprojections. A comparative study with two existing methods, and experimentation on simulated and real texture images is given.