Pedestrian Detection Algorithms using Shearlets PDF Download
Are you looking for read ebook online? Search for your book and save it on your Kindle device, PC, phones or tablets. Download Pedestrian Detection Algorithms using Shearlets PDF full book. Access full book title Pedestrian Detection Algorithms using Shearlets by Lienhard Pfeifer. Download full books in PDF and EPUB format.
Author: Lienhard Pfeifer
Publisher: Logos Verlag Berlin GmbH
ISBN: 3832548408
Category : Mathematics
Languages : en
Pages : 181
Get Book Here
Book Description
In this thesis, we investigate the applicability of the shearlet transform for the task of pedestrian detection. Due to the usage of in several emerging technologies, such as automated or autonomous vehicles, pedestrian detection has evolved into a key topic of research in the last decade. In this time period, a wealth of different algorithms has been developed. According to the current results on pedestrian detection benchmarks, the algorithms can be divided into two categories. First, application of hand-crafted image features and of a classifier trained on these features. Second, methods using Convolutional Neural Networks in which features are learned during the training phase. It is studied how both of these types of procedures can be further improved by the incorporation of shearlets, a framework for image analysis which has a comprehensive theoretical basis. To this end, we adapt the shearlet framework according to the requirements of the practical application of pedestrian detection algorithms. One main application area of pedestrian detection is located in the automotive domain. In this field, algorithms have to be runable on embedded devices. Therefore, we study the embedded realization of a pedestrian detection algorithm based on the shearlet transform.
Author: Lienhard Pfeifer
Publisher: Logos Verlag Berlin GmbH
ISBN: 3832548408
Category : Mathematics
Languages : en
Pages : 181
Get Book Here
Book Description
In this thesis, we investigate the applicability of the shearlet transform for the task of pedestrian detection. Due to the usage of in several emerging technologies, such as automated or autonomous vehicles, pedestrian detection has evolved into a key topic of research in the last decade. In this time period, a wealth of different algorithms has been developed. According to the current results on pedestrian detection benchmarks, the algorithms can be divided into two categories. First, application of hand-crafted image features and of a classifier trained on these features. Second, methods using Convolutional Neural Networks in which features are learned during the training phase. It is studied how both of these types of procedures can be further improved by the incorporation of shearlets, a framework for image analysis which has a comprehensive theoretical basis. To this end, we adapt the shearlet framework according to the requirements of the practical application of pedestrian detection algorithms. One main application area of pedestrian detection is located in the automotive domain. In this field, algorithms have to be runable on embedded devices. Therefore, we study the embedded realization of a pedestrian detection algorithm based on the shearlet transform.
Author: Ahmet Kapkic
Publisher:
ISBN:
Category : Computer vision
Languages : en
Pages : 51
Get Book Here
Book Description
Detecting pedestrians in public settings is a major research topic in both Computer Vision and Artificial Intelligence communities. It has found applications in a wide range of areas such as vehicle driving with autonomous control systems, video surveillance, and navigating robots, etc. Over the past decade, a great progress has been made in the development of efficient algorithms and the availability of large-scale data set, especially the advancement of Deep Learning method. In this thesis, the performance of a few state-of-the-art methods were evaluated by conducting empirical experiments with different settings and dataset configurations on pedestrian detection. The experiments were carried out using several Deep Learning models in the framework of both baseline and special configurations, including the Faster R-CNN, Mask R-CNN, and Cascade R-CNN methods. The experimental results show that the Mask R-CNN with a ResNet50 barebone yields the best performance in terms of its larger AP improvement and fewer resource requirement. This work provides a solid foundation upon which more sophisticated comparative studies can be conducted that utilize new algorithms/models and larger data set.
Author: Bastian Hartmann
Publisher: GRIN Verlag
ISBN: 3640981618
Category : Technology & Engineering
Languages : en
Pages : 89
Get Book Here
Book Description
Research Paper (undergraduate) from the year 2006 in the subject Electrotechnology, grade: 1,0, University Karlsruhe (TH), language: English, abstract: Avoiding collision accidents is becoming more and more an important topic in the research field of driver assistant systems. Especially for vision-based detection systems there are various approaches, which are built upon many different methods. This thesis deals with the avoidance of pedestrian accidents, caused by Blind Corner view problems. The presented approach comprises a pedestrian detection subsystem, which is part of a large camera system framework covering observation of the car environment. Based on a Blind Corner Camera and a neural network classification method, research in this thesis is focused on two aspects: detection improvement and danger level estimation. Since vision-based classification methods usually are still not able to yield perfect results, because of the complexity of this task, the detection result has to be improved by preprocessing and post processing. In this work, first, effects of image enhancement methods on detection are tested as preprocessing methods and, secondly, a new approach for a simple tracking and estimation strategy is presented, which improves detection in the way of a post processing method. Finally, information from tracking and prediction is used to estimate a danger level for pedestrians, which provides information about how collisionprone the current situations is.
Author: Oscar Lorente Corominas
Publisher:
ISBN:
Category :
Languages : en
Pages :
Get Book Here
Book Description
Pedestrian detection algorithms have always revolved around RGB scene information, but relying solely on it can be dangerous in situations where conventional cameras don't capture reality properly. For this reason, in recent years, many researchers have studied other alternatives that complement these existing techniques, such as the use of ultrasonic sensors or radars, that provide more reliable information in those situations. Another approach is to use LIDAR sensors, which map reality into point clouds using pulses of light. However, there are few studies that propose pedestrian detection techniques using only the data provided by a LIDAR. In this thesis, we explore this approach through the design and implementation of a pedestrian detection system in 3D point clouds. To do so, we train the PointNet++ point cloud classification network in order to demonstrate that the 3D geometric information of a scene is essential for the neural network to learn properly. Specifically, to carry out supervised training we need to generate a pedestrian and non-pedestrian ground truth in point clouds, so we have designed a semi-automatic labeling system based on the detection in RGB images and the subsequent transfer of these labels to the 3D domain. As a result, we train PointNet++ and test its performance on an outdoor dataset, obtaining outstanding results of up to 99.4% of accuracy and 98.6% of recall. With these results we are firmly corroborating the hypothesis stated in the thesis that 3D geometric information is essential for a neural network to learn to detect pedestrians in outdoor scenes. Not only that, we also surpass the results provided by a conventional detector in RGB images: YOLO, which provides a 48% of recall in the same dataset, thus proving that geometric information should not be an alternative in these systems, but a must.
Author: Philip Kelly
Publisher:
ISBN:
Category :
Languages : en
Pages : 0
Get Book Here
Book Description
Author:
Publisher:
ISBN:
Category : Computer vision
Languages : en
Pages : 232
Get Book Here
Book Description
There has been a significant increase in road accidents in the past few years due to vehicle driving popularization around the world. More than half of road fatalities are attributed to pedestrians. In order to reduce the number of pedestrian fatalities, many safety features have been developed, such as Advanced Driver Assistance Systems (ADAS). In ADAS technology, many on-vehicle sensors are used to detect the surrounding of the vehilce, and then this information is used to prevent accidents by sending warnings to the driver or taking over control of the vehicle, such as applying a brake. Vision-based detection algorithms are a widely used technology in ADAS for pedestrian detection due to the rich information they provide and their low cost compared to other sensors. Vision-based pedestrian detection is done in the following steps: image acquisition, candidate generation, feature extraction, classification, and real-time object tracking. This work focused on advancing the candidate generation step of the process. Generating potential pedestrian candidates from the input image is an important step in the detection system, and it has a significant impact in the detection accuracy and the algorithm run-time. Classifying a large number of unnecessary candidates increases the processing requirements and may result in false positives. There are many approaches for candidate generation. The basic way is the sliding window approach, where the whole image is scanned by a sliding window at multiple scales of its original size. Other approaches are selective, and they focus on certain regions of interest in the image for candidate generation. The stereo-vision approach for candidate generation is an example of a selective approach, where a 3-D map is constructed for the image view, and then candidates are generated from certain regions based on depth values. The common disadvantage in the current candidate generation approaches is the generation of a large number of unnecessary candidates, many of which are static background objects. Also, some of these approaches are computationally expensive. This dissertation introduces a new approach for pedestrian detection in a road infrastructure environment. The main idea of the proposed approach is to utilize the image frames provided by the previous vehicles that passed by a certain road section to more intelligently generate candidates. Vehicle-to-Infrastructure (V2I) communication is used to transmit image frames collected by vehicles for a certain location to the infrastructure database. The images are processed in the infrastructure for background modeling and moving object extraction. Candidates are generated from the moving object regions in the processed image. The proposed approach eliminates the candidates generated from static background objects, such as trees and buildings. The proposed model improves the detection accuracy by reducing the false positives and reducing the run-time of the detection algorithms. The system architecture of the proposed model is provided. The infrastructure algorithms for background modeling and pedestrian detection are implemented, and the results are analyzed and compared to an industry standard reference algorithm.
Author: Jennifer Vandoni
Publisher:
ISBN:
Category :
Languages : en
Pages : 0
Get Book Here
Book Description
This study deals with pedestrian detection in high- density crowds from a mono-camera system. The detections can be then used both to obtain robust density estimation, and to initialize a tracking algorithm. One of the most difficult challenges is that usual pedestrian detection methodologies do not scale well to high-density crowds, for reasons such as absence of background, high visual homogeneity, small size of the objects, and heavy occlusions. We cast the detection problem as a Multiple Classifier System (MCS), composed by two different ensembles of classifiers, the first one based on SVM (SVM-ensemble) and the second one based on CNN (CNN-ensemble), combined relying on the Belief Function Theory (BFT) to exploit their strengths for pixel-wise classification. SVM-ensemble is composed by several SVM detectors based on different gradient, texture and orientation descriptors, able to tackle the problem from different perspectives. BFT allows us to take into account the imprecision in addition to the uncertainty value provided by each classifier, which we consider coming from possible errors in the calibration procedure and from pixel neighbor's heterogeneity in the image space. However, scarcity of labeled data for specific dense crowd contexts reflects in the impossibility to obtain robust training and validation sets. By exploiting belief functions directly derived from the classifiers' combination, we propose an evidential Query-by-Committee (QBC) active learning algorithm to automatically select the most informative training samples. On the other side, we explore deep learning techniques by casting the problem as a segmentation task with soft labels, with a fully convolutional network designed to recover small objects thanks to a tailored use of dilated convolutions. In order to obtain a pixel-wise measure of reliability about the network's predictions, we create a CNN- ensemble by means of dropout at inference time, and we combine the different obtained realizations in the context of BFT. Finally, we show that the output map given by the MCS can be employed to perform people counting. We propose an evaluation method that can be applied at every scale, providing also uncertainty bounds on the estimated density.
Author: Tianyu Zuo
Publisher:
ISBN:
Category : University of Ottawa theses
Languages : en
Pages :
Get Book Here
Book Description
In this thesis, a novel Pedestrian Protection System (PPS), composed of the Pedestrian Detection System (PDS) and the Pedestrian Tracking System (PTS), was proposed. The PPS is a supplementary application for the Advanced Driver Assistance System, which is used to avoid collisions between vehicles and pedestrians. The Pedestrian Detection System (PDS) is used to detect pedestrians from near to far ranges with the feature-classi er-based detection method (HOG + SVM). To achieve pedestrian detection from near to far ranges, a novel structure was proposed. The structure of our PDS consists of two cameras (called CS and CL separately). The CS is equipped with a short focal length lens to detect pedestrians in near-to-mid range; and, the CL is equipped with a long focal length lens to detect pedestrians in mid-to-far range. To accelerate the processing speed of pedestrian detection, the parallel computing capacity of GPU was utilized in the PDS. The synchronization algorithm is also introduced to synchronize the detection results of CS and CL. Based on the novel pedestrian detection structure, the detection process can reach a distance which is more than 130 meters away without decreasing detection accuracy. The detection range can be extended more than 100 meters without decreasing the processing speed of pedestrian detection. Afterwards, an algorithm to eliminate duplicate detection results is proposed to improve the detection accuracy. The Pedestrian Tracking System (PTS) is applied following the Pedestrian Detection System. The PTS is used to track the movement trajectory of pedestrians and to predict the future motion and movement direction. A C + + class (called pedestrianTracking class, which is short for PTC) was generated to operate the tracking process for every detected pedestrian. The Kalman lter is the main algorithm inside the PTC. During the operation of PPS, the nal detection results of each frame from PDS will be transmitted to the PTS to enable the tracking process. The new detection results will be used to update the existing tracking results in the PTS. Moreover, if there is a newly detected pedestrian, a new process will be generated to track the pedestrian in the PTS. Based on the tracking results in PTS, the movement trajectory of pedestrians can be obtained and their future motion and movement direction can be predicted. Two kinds of alerts are generated based on the predictions: warning alert and dangerous alert. These two alerts represent di erent situations; and, they will alert drivers to the upcoming situations. Based on the predictions and alerts, the collisions can be prevented e ectively. The safety of pedestrians can be guaranteed.
Author: Keshava Grama
Publisher:
ISBN:
Category :
Languages : en
Pages :
Get Book Here
Book Description
The primary goal of this thesis is to present a robust low compute cost pedestrian tracking system for use with thermal infra-red images. Pedestrian tracking employs two distinct image analysis tasks, pedestrian detection and path tracking. This thesis will focus on benchmarking existing pedestrian tracking systems and using this to evaluate the proposed pedestrian detection and path tracking algorithm. The first part of the thesis describes the imaging system and the image dataset collected for evaluating pedestrian detection and tracking algorithms. The texture content of the images from the imaging system are evaluated using fourier maps following this the locations at which the dataset was collected are described. The second part of the thesis focuses on the detection and tracking system. To evaluate the performance of the tracking system, a time per target metric is described and is shown to work with existing tracking systems. A new pedestrian aspect ratio based pedestrian detection algorithm is proposed based on a binary matrix dynamically constrained using potential target edges. Results show that the proposed algorithm is effective at detecting pedestrians in infrared images while being less resource intensive as existing algorithms. The tracking system proposed uses deformable, dynamically updated codebook templates to track pedestrians in an infrared image sequence. Results show that this tracker performs as well as existing tracking systems in terms of accuracy, but requires fewer resources.
Author: Nicola Pellicanò
Publisher:
ISBN:
Category :
Languages : en
Pages : 0
Get Book Here
Book Description
Pedestrian detection and tracking have become important fields in Computer Vision research, due to their implications for many applications, e.g. surveillance, autonomous cars, robotics. Pedestrian detection in high density crowds is a natural extension of such research body. The ability to track each pedestrian independently in a dense crowd has multiple applications: study of human social behavior under high densities; detection of anomalies; large event infrastructure planning. On the other hand, high density crowds introduce novel problems to the detection task. First, clutter and occlusion problems are taken to the extreme, so that only heads are visible, and they are not easily separable from the moving background. Second, heads are usually small (they have a diameter of typically less than ten pixels) and with little or no textures. This comes out from two independent constraints, the need of one camera to have a field of view as high as possible, and the need of anonymization, i.e. the pedestrians must be not identifiable because of privacy concerns.In this work we develop a complete framework in order to handle the pedestrian detection and tracking problems under the presence of the novel difficulties that they introduce, by using multiple cameras, in order to implicitly handle the high occlusion issues.As a first contribution, we propose a robust method for camera pose estimation in surveillance environments. We handle problems as high distances between cameras, large perspective variations, and scarcity of matching information, by exploiting an entire video stream to perform the calibration, in such a way that it exhibits fast convergence to a good solution. Moreover, we are concerned not only with a global fitness of the solution, but also with reaching low local errors.As a second contribution, we propose an unsupervised multiple camera detection method which exploits the visual consistency of pixels between multiple views in order to estimate the presence of a pedestrian. After a fully automatic metric registration of the scene, one is capable of jointly estimating the presence of a pedestrian and its height, allowing for the projection of detections on a common ground plane, and thus allowing for 3D tracking, which can be much more robust with respect to image space based tracking.In the third part, we study different methods in order to perform supervised pedestrian detection on single views. Specifically, we aim to build a dense pedestrian segmentation of the scene starting from spatially imprecise labeling of data, i.e. heads centers instead of full head contours, since their extraction is unfeasible in a dense crowd. Most notably, deep architectures for semantic segmentation are studied and adapted to the problem of small head detection in cluttered environments.As last but not least contribution, we propose a novel framework in order to perform efficient information fusion in 2D spaces. The final aim is to perform multiple sensor fusion (supervised detectors on each view, and an unsupervised detector on multiple views) at ground plane level, that is, thus, our discernment frame. Since the space complexity of such discernment frame is very large, we propose an efficient compound hypothesis representation which has been shown to be invariant to the scale of the search space. Through such representation, we are capable of defining efficient basic operators and combination rules of Belief Function Theory. Furthermore, we propose a complementary graph based description of the relationships between compound hypotheses (i.e. intersections and inclusion), in order to perform efficient algorithms for, e.g. high level decision making.Finally, we demonstrate our information fusion approach both at a spatial level, i.e. between detectors of different natures, and at a temporal level, by performing evidential tracking of pedestrians on real large scale scenes in sparse and dense conditions.
