Pedestrian Detection

Visual object analysis researchers are increasingly experimenting with video, because it is expected that motion cues should help with detection, recognition, and other analysis tasks. This paper presents the Cambridge-driving Labeled Video Database (CamVid) as the first collection of videos with object class semantic labels, complete with metadata. The database provides ground truth labels that associate each pixel with one of 32 semantic classes. The database addresses the need for experimental data to quantitatively evaluate emerging algorithms. While most videos are filmed with fixed-position CCTV-style cameras, our data was captured from the perspective of a driving automobile. The driving scenario increases the number and heterogeneity of the observed object classes. Over 10 min of high quality 30 Hz footage is being provided, with corresponding semantically labeled images at 1 Hz and in part, 15 Hz. The CamVid Database offers four contributions that are relevant to object analysis researchers. First, the per-pixel semantic segmentation of over 700 images was specified manually, and was then inspected and confirmed by a second person for accuracy. Second, the high-quality and large resolution color video images in the database represent valuable extended duration digitized footage to those interested in driving scenarios or ego-motion. Third, we filmed calibration sequences for the camera color response and intrinsics, and computed a 3D camera pose for each frame in the sequences. Finally, in support of expanding this or other databases, we present custom-made labeling software for assisting users who wish to paint precise class-labels for other images and videos. We evaluate the relevance of the database by measuring the performance of an algorithm from each of three distinct domains: multi-class object recognition, pedestrian detection, and label propagation.

Advanced driver assistance systems (ADASs), and particularly pedestrian protection systems (PPSs), have become an active research area aimed at improving traffic safety. The major challenge of PPSs is the development of reliable on-board pedestrian detection systems. Due to the varying appearance of pedestrians (e.g., different clothes, changing size, aspect ratio, and dynamic shape) and the unstructured environment, it is very difficult to cope with the demanded robustness of this kind of system. Two problems arising in this research area are the lack of public benchmarks and the difficulty in reproducing many of the proposed methods, which makes it difficult to compare the approaches. As a result, surveying the literature by enumerating the proposals one-after-another is not the most useful way to provide a comparative point of view. Accordingly, we present a more convenient strategy to survey the different approaches. We divide the problem of detecting pedestrians from images into different processing steps, each with attached responsibilities. Then, the different proposed methods are analyzed and classified with respect to each processing stage, favoring a comparative viewpoint. Finally, discussion of the important topics is presented, putting special emphasis on the future needs and challenges.

Real-time detection of objects is receiving growing attention. The pedestrian is the most critical object that needs to be detecting and tracking by autonomous vehicles. The major challenges to this mission are caused by the difference in objects like pedestrians in age, gender, clothing, lighting, backgrounds, and occlusion. This paper starts with a brief introduction of problem-related to pedestrians, objects detection framework and Neural Networks Algorithms, and Real-Time Systems. And we focus on pedestrians as moving objects that need to detect, track, and solve problems related to computer vision. And based on our study we present a suggested solution for solving problems related to Pedestrians Detection in real-time particular. These techniques aim to be used in many applications such as Autonomous Vehicles, and Advanced Driver Assistance Systems (ADAS).

During the next decade, on-board pedestrian detection systems will play a key role in the challenge of increasing traffic safety. The main target of these systems, to detect pedestrians in urban scenarios, implies overcoming difficulties like processing outdoor scenes from a mobile platform and searching for aspect-changing objects in cluttered environments. This makes such systems combine techniques in the state-of-the-art Computer Vision. In this paper we present a three module system based on both 2D and 3D cues. The first module uses 3D information to estimate the road plane parameters and thus select a coherent set of regions of interest (ROIs) to be further analyzed. The second module uses Real AdaBoost and a combined set of Haar wavelets and edge orientation histograms to classify the incoming ROIs as pedestrian or non-pedestrian. The final module loops again with the 3D cue in order to verify the classified ROIs and with the 2D in order to refine the final results. According to the results, the integration of the proposed techniques gives rise to a promising system.

In this paper, we address the problem of multiperson tracking in busy pedestrian zones using a stereo rig mounted on a mobile platform. The complexity of the problem calls for an integrated solution that extracts as much visual information as possible and combines it through cognitive feedback cycles. We propose such an approach, which jointly estimates camera position, stereo depth, object detection, and tracking. The interplay between those components is represented by a graphical model. Since the model has to incorporate object-object interactions and temporal links to past frames, direct inference is intractable. We, therefore, propose a twostage procedure: for each frame, we first solve a simplified version of the model (disregarding interactions and temporal continuity) to estimate the scene geometry and an overcomplete set of object detections. Conditioned on these results, we then address object interactions, tracking, and prediction in a second step. The approach is experimentally evaluated on several long and difficult video sequences from busy inner-city locations. Our results show that the proposed integration makes it possible to deliver robust tracking performance in scenes of realistic complexity.

This experiment compared pedestrian detection using far-infrared (FIR) and near-infrared (NIR) night vision systems, combined with automatic warnings at one of two distances or no warning at all. Sixteen subjects (eight younger than 30 years and eight older than 60 years) pressed a button as soon as they saw a pedestrian on a night vision system in the center console of a vehicle simulator. In addition, they performed a concurrent simulated steering task that required almost continuous viewing of the forward scene, similar to real driving.

Penelitian ini mencoba menerapkan teknik
pengurangan latar belakang (background subtraction) untuk
melakukan segmentasi obyek pada citra. Citra digital diperoleh
dari perekaman dengan menggunakan kamera genggam pada
suatu area yang sering dilalui oleh pejalan kaki di dalam gedung.
Obyek pejalan kaki dipisahkan dari citra latar belakang dengan
melakukan pengurangan citra sederhana pada area RGB dan
grayscale. Keberadaan noise, bayangan dan ghost dapat
dihilangkan dengan proses penebalan (dilatasi) pada citra.
Klasifikasi sederhana dilakukan berdasarkan bentuk dari obyek.
Penelitian ini dapat mendeteksi obyek pejalan kaki dengan
cukup baik hanya dengan menggunakan teknik seleksi
berdasarkan ukuran dari obyek

This paper describes the recent research on the enhancement of pedestrian safety to help develop a better understanding of the nature, issues, approaches, and challenges surrounding the problem. It presents a comprehensive review of research efforts underway dealing with pedestrian safety and collision avoidance. The importance of pedestrian protection is emphasized in a global context, discussing the research programs and efforts in various countries. Pedestrian safety measures, including infrastructure enhancements and passive safety features in vehicles, are described, followed by a systematic description of active safety systems based on pedestrian detection using sensors in vehicle and infrastructure. The pedestrian detection approaches are classified according to various criteria such as the type and configuration of sensors, as well as the video cues and classifiers used in detection algorithms. It is noted that collision avoidance not only requires detection of pedestrians but also requires collision prediction using pedestrian dynamics and behavior analysis. Hence, this paper includes research dealing with probabilistic modeling of pedestrian behavior for predicting collisions between pedestrians and vehicles.

In this paper, we propose an approach for fast pedestrian detection in images. Inspired by the histogram of oriented gradient (HOG) features, a set of multi-scale orientation (MSO) features are proposed as the feature representation. The features are extracted on square image blocks of various sizes (called units), containing coarse and fine features in which coarse ones are the unit orientations and fine ones are the pixel orientation histograms of the unit. A cascade of Adaboost is employed to train classifiers on the coarse features, aiming to high detection speed. A greedy searching algorithm is employed to select fine features, which are input into SVMs to train the fine classifiers, aiming to high detection accuracy. Experiments report that our approach obtains state-of-art results with 12.4 times faster than the SVM+HOG method.

Objects recognition in image is one of the most difficult problems in computer vision. It is also an important step for the implementation of several existing applications that require highlevel image interpretation. Therefore, there is a growing interest in this research area during the last years. In this paper, we present an algorithm for human detection and recognition in realtime, from images taken by a CCD camera mounted on a car-like mobile robot. The proposed technique is based on Histograms of Oriented Gradient (HOG) and SVM classifier. The implementation of our detector has provided good results, and can be used in robotics tasks.

Pedestrian safety is a primary concern in transportation, especially in an urban environment. A variety of technologies, either infrastructure-based or vehicle-based, have evolved and emerged in recent years that offer promising prospects for detecting pedestrians. The sensing ...

This work was performed as part of the California PATH Program of the University of California, in cooperation with the State of California Business, Transportation, and Housing Agency, Department of Transportation, and the United States Department of ...

In this paper, we present a real-time pedestrian detection system that has been trained using a virtual environment. This is a very popular topic of research having endless practical applications and recently , there was an increasing interest in deep learning architectures for performing such a task. However, the availability of large labeled datasets is a key point for an effective train of such algorithms. For this reason, in this work, we introduced ViPeD, a new synthetically generated set of images extracted from a realistic 3D video game where the labels can be automatically generated exploiting 2D pedestrian positions extracted from the graphics engine. We exploited this new synthetic dataset fine-tuning a state-of-the-art computationally efficient Convolutional Neural Network (CNN). A preliminary experimental evaluation, compared to the performance of other existing approaches trained on real-world images , shows encouraging results.

Collision avoidance is one of the most difficult and challenging automatic driving operations in the domain of intelligent vehicles. In emergency situations, human drivers are more likely to brake than to steer, although the optimal maneuver would, more frequently, be steering alone. This statement suggests the use of automatic steering as a promising solution to avoid accidents in the future. The objective of this paper is to provide a collision avoidance system (CAS) for autonomous vehicles, focusing on pedestrian collision avoidance. The detection component involves a stereo-vision-based pedestrian detection system that provides suitable measurements of the time to collision. The collision avoidance maneuver is performed using fuzzy controllers for the actuators that mimic human behavior and reactions, along with a high-precision Global Positioning System (GPS), which provides the information needed for the autonomous navigation. The proposed system is evaluated in two steps. First, drivers' behavior and sensor accuracy are studied in experiments carried out by manual driving. This study will be used to define the parameters of the second step, in which automatic pedestrian collision avoidance is carried out at speeds of up to 30 km/h. The performed field tests provided encouraging results and proved the viability of the proposed approach.

Sokaklar kentteki hareketliliği sağlayan en küçük birimlerdir. İnsanı merkeze alan ve herkes için eşit erişim hakkı sağlayan bir sokak ağı, kenti daha yaşanabilir kılmaktadır. Güvenli ve erişilebilir bir sokak ağı ise kentsel yaşam kalitesini yükseltmektedir. Bugün gelişmiş birçok kent, vizyonunu, bisiklet ve yaya önceliğine sahip, sosyal etkileşime olanak sağlayan, güvenli sokaklar üzerinden kurgulamaktadır. Bir sokağın fiziksel ve sosyal kalitesini arttıran faktörlerin neler olduğu, tasarım öğelerinin ne şekilde belirleneceği, iyileştirme hamleleri için nasıl yöntem izleneceği yerel yönetimlerin gündemini meşgul eden önemli bir konudur. Günümüzde otomobil odaklı ulaşımın uygulamaları nedeniyle kentlerde otomobil dışındaki ulaşım biçimleri yok sayılmaktadır. Bu nedenle kentlerde yaya, engelli veya bisikletli olarak var olmak tehlikeye girmektedir. Her gün meydana gelen trafik kazaları, yollardaki altyapının yetersizliğinden doğan problemler bu durumun aciliyetine dikkat çekmektedir.  
“Yaya Dostu Kent” araştırması, kentlerimizde yaya, engelli, genç, yaşlı, bisikletli, kısacası her gruptan kentlinin güvenli ve konforlu bir şekilde erişimini amaçlamaktadır. Araştırma, sokak sakinlerinde farkındalık yaratma ve erişilebilir mahalle yaratmaya yönelik katılımcı araştırma süreci yürütmeyi hedeflemektedir.  Bu amaç doğrultusunda Esenler’de Sokak Denetimi yöntemi uygulanmış ve elde edilen sonuçlar değerlendirilerek öneriler geliştirilmiştir.

Advanced driver assistance systems (ADASs), and particularly pedestrian protection systems (PPSs), have become an active research area aimed at improving traffic safety. The major challenge of PPSs is the development of reliable on-board pedestrian detection systems. Due to the varying appearance of pedestrians (eg, different clothes, changing size, aspect ratio, and dynamic shape) and the unstructured environment, it is very difficult to cope with the demanded robustness of this kind of system.

Pedestrian detection is an important field in computer vision with applications in surveillance, robotics and driver assistance systems. The quality of such systems can be improved by the simultaneous use of different sensors. This paper proposes three different fusion techniques to combine the advantages of two vision sensors -a far-infrared (FIR) and a visible light camera. Different fusion methods taken from various levels of information representation are briefly described and finally compared regarding the results of the pedestrian classification.

This paper describes a vision-based pedestrian detection system for robots, and autonomous vehicles. For that purpose the Haar-like features were used to discriminate pedestrians. Those features were used as input in a learning algorithm, based on AdaBoost, which selects a small number of critical visual features from a larger set and yields an extremely efficient classifier. The proposed system can run in real-time applications achieving good detection rates.

In this paper we describe a fully integrated system for detecting, localizing, and tracking pedestrians from a moving vehicle. The system can reliably detect upright pedestrians to a range of 40 m in lightly cluttered urban environments. The system uses range data from stereo vision to segment the scene into regions of interest, from which shape features are extracted and used to classify pedestrians. The regions are tracked using shape and appearance features. Tracking is used to temporally filter classifications to improve performance and to estimate the velocity of pedestrians for use in path planning. The end-toend system runs at 5 Hz on 11024 1 768 imagery using a standard 2.4 GHz Intel Core 2 Quad processor, and has been integrated and tested on multiple ground vehicles and environments. We show performance on a diverse set of datasets with groundtruth in outdoor environments with varying degrees of pedestrian density and clutter. In highly cluttered urban environments, the detection rates are on a par with state-of-the-art but significantly slower systems.

This paper describes a stereo-vision-based candidate selection method for pedestrian detection from a moving vehicle. Non-dense 3D maps are computed by using epipolar geometry and a robust correlation process. Non-flat road assumption is used for correcting pitch angle variations. Thus, non obstacle points can be easily removed since they lay on the road. Generic obstacles are selected by using Subtractive Clustering algorithm in a 3D space with an adaptive radius. This clustering technique can be configurable for different types of obstacles. An optimal configuration for pedestrian detection is presented in this work.

This paper presents pedestrian detection algorithm on labeled depth data which is obtained from road scenes. Our approach computes feature responses for head and legs of human body using depth and label data. And then, it detects pedestrians by removing edges and partitioning a bipartite graph of head and leg response blobs using prior knowledge about human body. In the experiments, the proposed algorithm produces better result compared to the method which uses histogram of gradient feature and the ground plane for road scenes.

This paper describes an improved stereo vision system for anticipated detection of car-to-pedestrian accidents. An improvement of previous versions of the pedestrian dection system is achieved by compensation of the cameras pitch angle, since it results in higher accuracy in the location of the ground plane and more accurate depth measurements. The pedestrian detection system has been applied to collision avoidance and mitigation. Collision avoidance is carried out by means of deceleration strategies, whenever the accident is evitable. Likewise, collision mitigation is accomplished by activating an active hood system. For that purpose, the system has been mounted and tested on two different prototype cars and tested on private circuits using dummies.

The paper focuses on motion-based information extraction from video imagesequences. A novel method is introduced which can reliably detect walking human figures contained in such images. The method works with spatiotemporal input information to detect and classify the patterns typical of human movement. Our algorithm consists of easy-to-optimise operations, which in practical applications is an important factor. The paper presents a new information-extraction and temporal-tracking method based on a simplified version of the symmetry which is characteristic for the legs of a walking person. These spatio-temporal traces are labelled by kernel Fisher discriminant analysis. With this use of temporal tracking and non-linear classification we have achieved pedestrian detection from real-life images with a correct classification rate of 96.5%.

Automated methods are commonly used to count motorized vehicles, but are not frequently used to count pedestrians. This is because the automated technologies available to count pedestrians are not very developed, and their effectiveness has not been widely ...

The last few decades witnessed the birth and growth of a new sensibility to transportation efficiency. In particular, the need for efficient and improved people and goods mobility pushed researchers to address the problem of intelligent transportation systems. This paper surveys the most advanced approaches to the (partial) customization of road following task, using on-board systems based on artificial vision. The functionalities of lane detection, obstacle detection and pedestrian detection are described and classified, and their possible application on future road vehicles is discussed.

This paper presents a real-time and simple segmentation scheme for pedestrian detection system at night time using NIR camera. The system consists of an NIR camera used to capture images at night time. The obtained images are sent through segmentation and tracking blocks to detect the pedestrians. A new vertical edge detection method is used to detect edges due to pedestrians and the edges, thus formed, are combined to form candidate blocks. The blocks are further segmented using profiling to obtain tight bound boxes. The candidate blocks are rejected if the height obtained does not fall in the limits of the estimated height (obtained from camera settings). Two different schemes based on template matching and segmentation of the smaller image blocks are experimented for tracking purpose. The experiments performed show that the algorithm can be used for real-time applications.

Pedestrian safety is B primary traffic issue in urban environment. The use of modern sensing technologies to improve pedestrian safety has remained an active research topic for years. A variety of sensing technologies have been developed for pedestrian detection. The application of pedestrian detection on transit vehicle platforms is desirable and feasible in the near future. In this paper, potential sensing technologies are first reviewed for their advantages and limitations. Several sensors are then chosen for further experhntal testing and evaluation. A reliable sensing system will require a combination of multiple sensors to deal with near-range in stationary conditions and longer-range detection in moving conditions. An approach of vehicle-infrastructure integrated solution is suggested for the pedestrian detection in transit bus application.

In this paper, we present a real-time pedestrian detection system that has been trained using a virtual environment. This is a very popular topic of research having endless practical applications and recently, there was an increasing interest in deep learning architectures for performing such a task. However, the availability of large labeled datasets is a key point for an effective train of such algorithms. For this reason, in this work, we introduced ViPeD, a new synthetically generated set of images extracted from a realistic 3D video game where the labels can be automatically generated exploiting 2D pedestrian positions extracted from the graphics engine. We exploited this new synthetic dataset fine-tuning a state-of-the-art computationally efficient Convolutional Neural Network (CNN). A preliminary experimental evaluation, compared to the performance of other existing approaches trained on real-world images, shows encouraging results.

Recently, deep learning methods, mostly algorithms based on Deep Convolutional Neural Networks (DCNNs) have yielded great results on pedestrian detection. Algorithms based on DCNNs spontaneously learn features in a supervised manner and are able to learn qualified high level feature representations to detect pedestrian. In this paper, we first review a number of popular DCNN-based training approaches along with their recent extensions. We then briefly describe recent algorithms based on these approaches. Also, we accentuate recent contributions and main challenges of DCNNs in detecting pedestrian. We analyze deep pedestrian detection algorithms from training approach, categorization, and DCNN model points of view, and ultimately propose a new deep architecture and training approach for deep pedestrian detection. The experimental results show that the proposed DCNN and training approach, achieve more accurate rate detection than the previously reported architectures and training approaches.

Visible light communication (VLC) systems are promising candidates for future indoor access and peer-to-peer networks. The performance of these systems, however, is vulnerable to line of sight (LOS) link blockage due to objects inside the room. Considering pedestrians as the most common VLC links blocking obstacles, we develop a probabilistic passive pedestrian detection and localization method. Our method takes advantage of the blockage status of VLC LOS links between the user equipment (UE) and transceivers on the ceiling to passively detect a single pedestrian, modeled as a cylinder with a random radius. The VLC network gathers the blockage status and computes the geometry of the LOS link graph through a cooperative scheme between VLC device-equipped users inside the room. We also develop a mathematical framework to obtain an optimum solution for estimating the location and size of the object and conclude with a sub-optimum estimation by simplifying the problem to a quadratic programming approach. Simulation results show that using a 5 × 5 grid of transceivers on the ceiling and as few as eight UEs, the root-mean-squared error in estimating the center and radius of the object can be less than 5 cm and 3 cm, respectively.

This paper presents an efficient technique for real time estimation of on-board stereo vision system pose. The whole process is performed in the Euclidean space and consists of two stages. Initially, a compact representation of the original 3D data points is computed. Then, a RANSAC based least squares approach is used for fitting a plane to the 3D road points. Fast RANSAC fitting is obtained by selecting points according to a probability distribution function that takes into account the density of points at a given depth. Finally, stereo camera position and orientation-pose-is computed relative to the road plane. The proposed technique is intended to be used on driver assistance systems for applications such as obstacle or pedestrian detection. A real time performance is reached. Experimental results on several environments and comparisons with a previous work are presented.

In general, a distributed processing is not suitable for dealing with image data stream due to the network load problem caused by communications of frames. For this reason, image data stream processing has operated in just one node commonly. However, we need to process image data stream in a distributed environment in a big data era due to increase in quantity and quality of multimedia data. In this paper, we shall present a real-time pedestrian detection methodology in a distributed environment which processes image data stream in real-time on Apache Storm framework. It achieves sharp speed up by distributing frames onto several nodes called bolts, each of which processes different regions of image frames. Moreover, it can reduce the overhead caused by synchronization by computation bolts which returns only the processing results to the merging bolts.

In automotive design, the issue of safety remains a growing priority. Recently the focus has extended beyond the occupants of the vehicle and has turned towards other Vulnerable Road Users (VRU). Simple night vision systems have already become an important safety feature in modern high end automobiles. The next generation of advanced driver assistance systems will automate the detection of VRUs, to improve safety further by not distracting the driver's attention from the road ahead, and even identifying dangerous situations where the driver may not. This paper presents a review of the state of the art image processing techniques for automatic detection and classification of VRUs in automotive far infrared imagery.

Collision avoidance is one of the most difficult and challenging automatic driving operations in the domain of intelligent vehicles. In emergency situations, human drivers are more likely to brake than to steer, although the optimal maneuver would, more frequently, be steering alone. This statement suggests the use of automatic steering as a promising solution to avoid accidents in the future. The objective of this paper is to provide a collision avoidance system (CAS) for autonomous vehicles, focusing on pedestrian collision avoidance. The detection component involves a stereo-vision-based pedestrian detection system that provides suitable measurements of the time to collision. The collision avoidance maneuver is performed using fuzzy controllers for the actuators that mimic human behavior and reactions, along with a high-precision Global Positioning System (GPS), which provides the information needed for the autonomous navigation. The proposed system is evaluated in two steps. First, drivers' behavior and sensor accuracy are studied in experiments carried out by manual driving. This study will be used to define the parameters of the second step, in which automatic pedestrian collision avoidance is carried out at speeds of up to 30 km/h. The performed field tests provided encouraging results and proved the viability of the proposed approach.

This article presents a tetra-vision (4 cameras) system for the detection of pedestrians by the means of the simultaneous use of one far infra-red and one visible cameras stereo pairs. The main idea is to exploit both the advantages of far infra-red and visible cameras trying at the same time to benefit from the use of each system. Initially, the two stereo flows are independently processed, then the results are fused together. The final result of this low-level processing is a list of obstacles that have a shape and a size compatible with the presence of a potential pedestrian. In addition, the system is able to remove the background from the detected obstacles to simplify a possible further high level processing.

Pedestrians are the most vulnerable participants to urban traffic. The first step toward protecting pedestrians is to reliably detect them. We present a new approach for standing and walking pedestrian detection, in urban traffic conditions, using greyscale stereo cameras mounted on board a vehicle. Our system uses pattern matching and motion for pedestrian detection. Both 2-D image intensity information and 3-D dense stereo information are used for classification. The 3-D data is used for effective pedestrian hypothesis generation, scale and depth estimation and 2-D model selection. The scaled models are matched against the selected hypothesis using a high performance, elastic matching, based on the Chamfer distance. Kalman filtering is used to track detected pedestrians. A subsequent validation, based on the motion field's variance and periodicity of tracked, walking pedestrians is used to eliminate false positives.

This paper describes a system for pedestrian detection in infrared images, which has been implemented on an experimental vehicle equipped with an infrared camera. The proposed system has been tested in many situations and has proven to be efficient and with a very low false-positive rate. It is based on a multiresolution localization of warm symmetrical objects with specific size and aspect ratio; anyway, because road infrastructures and other road participants may also have such characteristics, a set of matched filters is included in order to reduce false detections. A final validation process, based on human shape's morphological characteristics, is used to build the list of pedestrian appearing in the scene. Neither temporal correlation nor motion cues are used in this first part of the project: the processing is based on the analysis of single frames only.

Pedestrian detection is essential to avoid dangerous traffic situations. In this paper, we present a fast and robust algorithm for detecting pedestrians in a cluttered scene from a pair of moving cameras. This is achieved through stereo-based segmentation and neural network-based recognition. The algorithm includes three steps. First, we segment the image into sub-image object candidates using disparities discontinuity. Second, we merge and split the sub-image object candidates into sub-images that satisfy pedestrian size and shape constrains. Third, we use intensity gradients of the candidate sub-images as input to a trained neural network for pedestrian recognition. The experiments on a large number of urban street scenes demonstrate that the proposed algorithm: 1) can detect pedestrians in various poses, shapes, sizes, clothing, and occlusion status; 2) runs in real-time; and 3) is robust to illumination and background changes.

The detection of pedestrians in real-world scenes is a daunting task, especially in crowded situations. Our experience over the last years has shown that active shape models (ASM) can contribute significantly to a robust pedestrian detection system. The paper starts with an overview of shape model approaches, it then explains our approach which builds on top of Eigenshape models which have been trained using real-world data. These models are placed over candidate regions and matched to image gradients using a scoring function which integrates i) point distribution, ii) local gradient orientations iii) local image gradient strengths. A matching and shape model update process is iteratively applied in order to fit the flexible models to the local image content. The weights of the scoring function have a significant impact on the ASM performance. We analyze different settings of scoring weights for gradient magnitude, relative orientation differences, distance between model and gradient in an experiment which uses real-world data. Although for only one pedestrian model in an image computation time is low, the number of necessary processing cycles which is needed to track many people in crowded scenes can become the bottleneck in a real-time application. We describe the measures which have been taken in order to improve the speed of the ASM implementation and make it real-time capable.

This paper describes a stereo-vision-based pedestrian detection system for intelligent transportation systems. The basic components of pedestrians are first located in the image and then combined with a SVM-based classifier. Generic obstacles are located using a subtractive clustering attention mechanism based on stereo vision. A by-components learning approach is proposed and different feature extraction methods are tested in order to better deal with pedestrian variability and justify what features are better to be learnt for pedestrian detection. Candidate selection mechanisms usually yield pedestrians with inaccurate bounding boxes. Then a decrease in detection rate takes place if the SVM classifier is trained only with well-fitted pedestrians. Using several off-line databases containing thousands of pedestrians samples the effect of bounding box accuracy is studied. A multi-candidate generation mechanism is also developed in order to enhance the single frame performance, decreasing the number of false positives due to inaccurate bounding boxes