A5083927606- Video Surveillance and Tracking Methods
- Advanced Vision and Imaging
- Autonomous Vehicle Technology and Safety
- Human Pose and Action Recognition
- Anomaly Detection Techniques and Applications
- Fire Detection and Safety Systems
- Automated Road and Building Extraction
- Gait Recognition and Analysis
- Vehicle License Plate Recognition
- Image and Object Detection Techniques
- Image Enhancement Techniques
- Robotics and Sensor-Based Localization
- Optical measurement and interference techniques
- Infrastructure Maintenance and Monitoring
- Industrial Vision Systems and Defect Detection
- Remote Sensing and LiDAR Applications
- Advanced X-ray and CT Imaging
- Gaze Tracking and Assistive Technology
- Radiomics and Machine Learning in Medical Imaging
- Transportation Safety and Impact Analysis
- Smart Materials for Construction
- Robot Manipulation and Learning
- Video Analysis and Summarization
- Ultrasound and Hyperthermia Applications
- Thyroid Cancer Diagnosis and Treatment
Benha University
2021
Vital Research
2007-2019
Toshiba (United States)
2016
National Research Institute of Astronomy and Geophysics
2011
University of Minnesota
1999-2009
Twin Cities Orthopedics
2002-2007
University of Minnesota System
1999
This paper presents algorithms for vision-based detection and classification of vehicles in monocular image sequences traffic scenes recorded by a stationary camera. Processing is done at three levels: raw images, region level, vehicle level. Vehicles are modeled as rectangular patches with certain dynamic behavior. The proposed method based on the establishment correspondences between regions vehicles, move through sequence. Experimental results from highway provided which demonstrate...
This paper presents a real-time system for pedestrian tracking in sequences of grayscale images acquired by stationary camera. The objective is to integrate this with traffic control application such as scheme at intersections. proposed approach can also be used detect and track humans front vehicles. Furthermore, the schemes employed detection several diverse objects interest (vehicles, bicycles, etc.) outputs spatio-temporal coordinates each during period scene. Processing done three...
The goal of this project is to monitor activities at traffic intersections for detecting/predicting situations that may lead accidents. Some the key elements robust intersection monitoring are camera calibration, motion tracking, incident detection, etc. In paper, we consider motion-tracking problem. A multilevel tracking approach using Kalman filter presented vehicles and pedestrians intersections. combines low-level image-based blob with high-level filtering position shape estimation. An...
Monitoring traffic intersections in real time and predicting possible collisions is an important first step towards building early collision-warning system. We present a vision-based system addressing this problem describe the practical adaptations necessary to achieve real-time performance. Innovative low-overhead collision-prediction algorithms (such as one using time-as-axis paradigm) are presented. The proposed was able perform successfully on videos of quarter-video graphics array (VGA)...
This paper presents a vision-based method to automatically detect individuals loitering about inner-city bus stops. Using stationary camera view of stop, pedestrians are segmented and tracked throughout the scene. The system takes snapshots when clean, nonobstructed pedestrian is found. then used classify individual images into database, using an appearance-based method. features correlate based on short-term biometrics, which changeable but stay valid for short periods time; this uses...
This work presents a method for detecting abandoned objects in real-world conditions. The presented here addresses the online and real time aspects of such systems, utilizes logic to differentiate between stationary people, is robust temporary occlusion potential objects. capacity not detect still people as major aspect that differentiates this from others literature. Results are on 3 hours 36 minutes footage over four videos representing both sparsely densely populated situations, also...
We address the problem of automatically learning layout a traffic intersection from trajectories vehicles obtained by vision tracking system. present similarity measure which is suitable for use with spectral clustering in problems that emphasize spatial distinctions between vehicle trajectories. The robustness method to small perturbations and its sensitivity choice parameters are evaluated using real-world data
This paper presents a novel technique for camera tampering detection. It is implemented in real-time and was developed use surveillance security applications. method identifies by detecting large differences between older frames of video more recent frames. A buffer incoming kept three different measures image dissimilarity are used to compare the After normalization, set conditions tested decide if has occurred. The effects adjusting internal parameters algorithm examined. performance this...
This project involves the use of a dashboard-mounted camera to monitor direction driver is looking. accomplished by using framework for processing video (FPV), developed at University Minnesota Osama Masoud. The monitoring software works first finding lips on driver's face color analysis. Then, skin can be sampled and region generated. largest holes in this are eyes. Once eyes found, finds darkest pixels marks these as pupils. uses relative positions pupils make statements about gaze...
This article presents a vision-based system for monitoring crowded urban scenes. The approach combines an effective detection scheme based on optical flow and background removal that can locate vehicles, individual pedestrians, crowds. phase is followed by the tracking tracks all detected entities. Traffic objects are not simply tracked but wealth of information (position, velocity, acceleration/deceleration, bounding rectangle, shape features) gathered about them also. Potential...
This paper deals with real-time image processing of crowded outdoor scenes the objective creating an effective traffic management system that monitors urban settings (urban intersections, streets after athletic events, etc.). The proposed can detect, track, and monitor both pedestrians (crowds) vehicles. We describe characteristics tracker is based on a new detection method. Initially, we produce motion estimation map. map then segmented analyzed in order to remove inherent noise focus...
The paper deals with the problem of classification human activities from video as one way performing activity monitoring. Our approach uses motion features that are computed very efficiently and subsequently projected into a lower dimension space where matching is performed. Each action represented manifold in this done by comparing these manifolds. To demonstrate effectiveness approach, it was used on large data set similar actions, each performed many different actors. Classification...
In this paper, we propose a multi-level shadow identification scheme which is generally applicable without restrictions on the number of light sources, illumination conditions, surface orientations, and object sizes. first level, use background segmentation technique to identify foreground regions include moving shadows. second step, pixel-based decisions are made by comparing current frame with model distinguish between shadows actual foreground. third result improved using blob-level...
We discuss some of the practical issues concerning use mixtures Gaussians for background segmentation in outdoor scenes, including choice parameters. Different covariance representations and their performance impact are examined. In addition, we propose a simple, yet efficient method coping with sudden global illumination changes based on smoothing brightness contrast over time. All discussed methods capable running real time at reasonable resolution current generation PCs.
This paper presents algorithms for vision-based detection and classification of vehicles in monocular image sequences traffic scenes recorded by a stationary camera. Processing is done at three levels: raw images, blob level vehicle level. Vehicles are modeled as rectangular patches with certain dynamic behavior. Kalman filtering used to estimate parameters. The proposed method based on the establishment correspondences among blobs vehicles, move through sequence. Experimental results from...
We propose a motion recognition strategy that represents videoclip as set of filtered images, each which encodes short period history. Given videoclips whose types are known, image classifier is built using support vector machines. In offline classification, the label test obtained by applying majority voting over its images. online most probable type action at an instance determined recent within sliding window. The effectiveness this was demonstrated on real datasets where were recorded...
This paper presents algorithms for vision-based monitoring of weaving sections. These have been developed the Minnesota Department Transportation in order to acquire data several sections Twin Cities area. Unlike commercially available systems, proposed can track and count vehicles as they change lanes. Furthermore, provide velocity direction each vehicle section. Experimental results from various under weather conditions are presented. The methods based on establishment correspondences...
The goal of tills project Is to develop a passive vision-based sensing system. system will be capable monitoring an intersection by observing the vehicle and pedestrian flow, predicting situations that might give rise accidents. A single camera looking at all from arbitrary position is used. However, for extended applications, multiple cameras needed. Some key elements are calibration, motion tracking, classification, giving collisions. In this paper, we focus on tracking. Motion...
This paper presents a real-time system for pedestrian tracking in sequences of grayscale images acquired by stationary CCD camera. The objective is to integrate this with control scheme intersections. outputs the spatio-temporal coordinates each during period scene. Processing done at three levels: raw images, blobs, and pedestrians. Our method models pedestrians as rectangular patches certain dynamic behavior. Kalman filtering used estimate parameters. was implemented on Datacube MaxVideo...
In this paper, we address the problem of recovering intrinsic and extrinsic parameters a camera or group cameras in setting overlooking traffic scene. Unlike many other settings, conventional calibration techniques are not applicable case. We present method that uses certain geometric primitives commonly found scenes order to recover parameters. These provide needed redundancy weighted depending on significance there corresponding image features. show experimentally these capable achieving...