We present a novel pedestrian representation based on a new model of pedestrian motion coupled with a geometric optimization method. The model of pedestrian motion seeks to capture the underlying physiological and psychological factors which give rise to the fundamental diagram — the phenomenon that pedestrian speed reduces as density increases. The optimization method computes collision-free v...

Pedestrian detection is a hot topic in computer vision and pattern recognition. Existing pedestrian detection methods face new challenges in the background of big data, e.g., heavy burdens on computing and memory. To solve these problems, in this paper, we propose a pedestrian detection framework based on incremental learning. Compared with existing pedestrian detection frameworks, it costs muc...

Pedestrian behavioral modeling is a topic that is receiving more and more attention in different areas of application such as panic analysis, building evacuation and surveillance systems. We propose a model based on discrete choice theory for pedestrian dynamic behavior. Our model is based on a local discretization of the space around each pedestrian, as opposed to other models which are contin...

We present a new method for obstacle and pedestrian detection with lower computation complexity and higher detection performance on a small target. By calculating stereo disparity according to the road parameters, our system efficiently detects objects above the ground. Experiments on over 10,000 images captured in urban areas demonstrate our method’s effectiveness. We also propose a multi-scal...

We present a novel framework for socially competent pedestrian navigation based on understanding pedestrians’ intentions and planning intent-expressive robot motion. We model pedestrians’ intentions as combinations of intended topological routes and intended destinations. The core of this approach is a novel topological representation of a pedestrian scene, based on braid groups. This represent...

Learning to recognize pedestrian attributes at far distance is a novel research topic in video surveillance scenarios where face and body close-shots are hardly available; instead, only far-view video frames of pedestrian are given. In this study, we present an alternative approach that exploits the context of neighboring pedestrian images for improved attribute inference compared to the conven...

We have introduced excluded volume effect, which is an important factor to model a realistic pedestrian queue, into queueing theory. The probability distributions of pedestrian number and pedestrian waiting time in a queue have been calculated exactly. Due to time needed to close up the queue, the mean number of pedestrians increases as pedestrian arrival probability (λ) and leaving probability...

This paper proposes a novel vision based multi-pedestrian tracking scheme in crowded scenes, which are very common in real-world applications. The major challenge of the multi-pedestrian tracking problem comes from complicated occlusions, cluttered or even changing background. We address these issues by creatively combining state-of-the-art pedestrian detectors and clustering algorithms. The co...

Designing pedestrian systems have been a challenge due to the lack of effective tools and methods to model the autonomous behavior of pedestrians. Pedestrians are one of the main players in traffic systems, and they also are their most common victims. The high incident of pedestrians’ accidents has led the World Health Organization (WHO) to set the reduction of traffic accidents as a priority f...

Studies on microscopic pedestrian requires large amounts of trajectory data from real-world pedestrian crowds. Such data collection, if done manually, needs tremendous effort and is very time consuming. Though many studies have asserted the possibility of automating this task using video cameras, we found that only a few have demonstrated good performance in very crowded situations or from a to...