Francesco Solera

Performance Measures and a Data Set for Multi-target, Multi-camera Tracking

To help accelerate progress in multi-target, multi-camera tracking systems, we present (i) a new pair of precision-recall measures of performance that treats errors of all types uniformly and emphasizes correct identification over sources of error; (ii) the largest fully-annotated and calibrated data set to date with more than 2 million frames of 1080 p, 60 fps video taken by 8 cameras observing more than 2,700 identities over 85 min; and (iii) a reference software system as a comparison baseline. We show that (i) our measures properly account for bottom-line identity match performance in the multi-camera setting; (ii) our data set poses realistic challenges to current trackers; and (iii) the performance of our system is comparable to the state of the art.

Learning to Divide and Conquer for Online Multi-target Tracking

Online Multiple Target Tracking (MTT) is often addressed within the tracking-by-detection paradigm. Detections are previously extracted independently in each frame and then objects trajectories are built by maximizing specifically designed coherence functions. Nevertheless, ambiguities arise in presence of occlusions or detection errors. In this paper we claim that the ambiguities in tracking could be solved by a selective use of the features, by working with more reliable features if possible and exploiting a deeper representation of the target only if necessary. To this end, we propose an online divide and conquer tracker for static camera scenes, which partitions the assignment problem in local subproblems and solves them by selectively choosing and combining the best features. The complete framework is cast as a structural learning task that unifies these phases and learns tracker parameters from examples. Experiments on two different datasets highlights a significant improvement of tracking performances (MOTA +10%) over the state of the art.

Structured learning for detection of social groups in crowd

Group detection in crowds will play a key role in future behavior analysis surveillance systems. In this work we build a new Structural SVM-based learning framework able to solve the group detection task by exploiting annotated video data to deduce a sociologically motivated distance measure founded on Halls proxemics and Grangers causality. We improve over state-of-the-art results even in the most crowded test scenarios, while keeping the classification time affordable for quasi-real time applications. A new scoring scheme specifically designed for the group detection task is also proposed.

From Ego to Nos-Vision: Detecting Social Relationships in First-Person Views

In this paper we present a novel approach to detect groups in ego-vision scenarios. People in the scene are tracked through the video sequence and their head pose and 3D location are estimated. Based on the concept of f-formation, we define with the orientation and distance an inherently social pairwise feature that describes the affinity of a pair of people in the scene. We apply a correlation clustering algorithm that merges pairs of people into socially related groups. Due to the very shifting nature of social interactions and the different meanings that orientations and distances can assume in different contexts, we learn the weight vector of the correlation clustering using Structural SVMs. We extensively test our approach on two publicly available datasets showing encouraging results when detecting groups from first-person camera views.

Tracking Social Groups Within and Across Cameras

We propose a method for tracking groups from single and multiple cameras with disjointed fields of view. Our formulation follows the tracking-by-detection paradigm in which groups are the atomic entities and are linked over time to form long and consistent trajectories. To this end, we formulate the problem as a supervised clustering problem in which a structural SVM classifier learns a similarity measure appropriate for group entities. Multicamera group tracking is handled inside the framework by adopting an orthogonal feature encoding that allows the classifier to learn inter- and intra-camera feature weights differently. Experiments were carried out on a novel annotated group tracking data set, the DukeMTMC-Groups data set. Since this is the first data set on the problem, it comes with the proposal of a suitable evaluation measure. Results of adopting learning for the task are encouraging, scoring a +15% improvement in F1 measure over a nonlearning-based clustering baseline. To the best of our knowledge, this is the first proposal of its kind dealing with multicamera group tracking.

Towards the evaluation of reproducible robustness in tracking-by-detection

Conventional experiments on MTT are built upon the belief that fixing the detections to different trackers is sufficient to obtain a fair comparison. In this work we argue how the true behavior of a tracker is exposed when evaluated by varying the input detections rather than by fixing them. We propose a systematic and reproducible protocol and a MATLAB toolbox for generating synthetic data starting from ground truth detections, a proper set of metrics to understand and compare trackers peculiarities and respective visualization solutions.

Learning where to attend like a human driver

Despite the advent of autonomous cars, its likely — at least in the near future — that human attention will still maintain a central role as a guarantee in terms of legal responsibility during the driving task. In this paper we study the dynamics of the drivers gaze and use it as a proxy to understand related attentional mechanisms. First, we build our analysis upon two questions: where and what the driver is looking at? Second, we model the drivers gaze by training a coarse-to-fine convolutional network on short sequences extracted from the DR(eye)VE dataset. Experimental comparison against different baselines reveal that the drivers gaze can indeed be learnt to some extent, despite i) being highly subjective and ii) having only one drivers gaze available for each sequence due to the irreproducibility of the scene. Eventually, we advocate for a new assisted driving paradigm which suggests to the driver, with no intervention, where she should focus her attention.

Learning to identify leaders in crowd

Leader identification is a crucial task in social analysis, crowd management and emergency planning. In this paper, we investigate a computational model for the individuation of leaders in crowded scenes. We deal with the lack of a formal definition of leadership by learning, in a supervised fashion, a metric space based exclusively on people spatiotemporal information. Based on Tardes work on crowd psychology, individuals are modeled as nodes of a directed graph and leaders inherits their relevance thanks to other members references. We note this is analogous to the way websites are ranked by the PageRank algorithm. During experiments, we observed different feature weights depending on the specific type of crowd, highlighting the impossibility to provide a unique interpretation of leadership. To our knowledge, this is the first attempt to study leader identification as a metric learning problem.

Social Groups Detection in Crowd through Shape-Augmented Structured Learning

Most of the behaviors people exhibit while being part of a crowd are social processes that tend to emerge among groups and as a consequence, detecting groups in crowds is becoming an important issue in modern behavior analysis. We propose a supervised correlation clustering technique that employs Structural SVM and a proxemic based feature to learn how to partition people trajectories in groups, by injecting in the model socially plausible shape configurations. By taking into account social groups patterns, the system is able to outperform state of the art methods on two publicly available benchmark sets of videos.

From Groups to Leaders and Back: Exploring Mutual Predictability Between Social Groups and Their Leaders.

Abstract Recently, social theories and empirical observations identified small groups and leaders as the basic elements which shape a crowd. This leads to an intermediate level of abstraction that is placed between the crowd as a flow of people, and the crowd as a collection of individuals. Consequently, automatic analysis of crowds in computer vision is also experiencing a shift in focus from individuals to groups and from small groups to their leaders. In this chapter, we present state-of-the-art solutions to the groups and leaders detection problem, which are able to account for physical factors as well as for sociological evidence observed over short time windows. The presented algorithms are framed as structured learning problems over the set of individual trajectories. However, the way trajectories are exploited to predict the structure of the crowd is not fixed but rather learned from recorded and annotated data, enabling the method to adapt these concepts to different scenarios, densities, cultures, and other unobservable complexities. Additionally, we investigate the relation between leaders and their groups and propose the first attempt to exploit leadership as prior knowledge for group detection.