Correlating Multiple 2D Vision Trackers for 3D Object Tracking on Construction Sites

Tracking methods have the potential to retrieve the spatial location of project related entities such as personnel and equipment at construction sites, which can facilitate several construction management tasks. Existing tracking methods are mainly based on Radio Frequency (RF) technologies and thus require manual deployment of tags. On construction sites with numerous entities, tags installation, maintenance and decommissioning become an issue since it increases the cost and time needed to implement these tracking methods. To address these limitations, this paper proposes an alternate 3D tracking method based on vision. It operates by tracking the designated object in 2D video frames and correlating the tracking results from multiple pre-calibrated views using epipolar geometry. The methodology presented in this paper has been implemented and tested on videos taken in controlled experimental conditions. Results are compared with the actual 3D positions to validate its performance.

Detection of Construction Workers in Video Frames for Automatic Initialization of Vision Trackers

Vision based tracking can provide the spatial location of construction entities such as equipment, workers, and materials in large scale, congested construction sites. It tracks entities in video streams by inferring their locations based on the entities’ visual features and motion histories. To initiate the process, it is necessary to determine the pixel areas corresponding to the construction entities to be tracked in the following consecutive video frames. In order to fully automate the process, an automated way of initialization is needed. This paper presents the method for construction worker detection which can automatically recognize and localize construction workers in video frames. The method first finds the foreground areas of moving objects using a background subtraction method. Within these foreground areas, construction workers are recognized based on the histogram of oriented gradients (HOG) and histogram of the HSV colors. HOG’s have proved to work effectively for detection of people, and the histogram of HSV colors helps differentiate between pedestrians and construction workers wearing safety vests. Preliminary experiments show that the proposed method has the potential to automate the initialization process of vision based tracking.

IREE: Automated Vision Tracking of Project Related Entities

The purpose of this supplemental project was to collect invaluable data from the large-scale construction sites of Egnatia Odos motorway needed to validate a novel automated vision-tracking method created under the parent grant. For this purpose, one US graduate and three US undergraduate students traveled to Greece for 4 months and worked together with 2 Greek graduate students of the local faculty collaborator. This team of students monitored project activities and scheduled data collection trips on a daily basis, setup a mobile video data collection lab on the back of a truck, and drove to various sites every day to collect hundreds of hours of video from multiple cameras on a large variety of activities ranging from soil excavation to bridge construction. The US students were underrepresented students from minority groups who had never visited a foreign country. As a result, this trip was a major life experience to them. They learned how to live in a non-English speaking country, communicate with Greek students, workers and engineers. They lead a project in a very unfamiliar environment, troubleshoot myriad problems that hampered their progress daily and, above all, how to collaborate effectively and efficiently with other cultures. They returned to the US more mature, with improved leadership and problem-solving skills and a wider perspective of their profession.

Construction Worker Detection in Video Frames for Initializing Vision Trackers

Monitoring the location of resources on large scale, congested, outdoor sites can be performed more efficiently with vision tracking, as this approach does not require any pre-tagging of resources. However, the greatest impediment to the use of vision tracking in this case is the lack of detection methods that are needed to automatically mark the resources of interest and initiate the tracking. This paper presents such a novel method for construction worker detection that localizes construction workers in video frames. The proposed method exploits motion, shape, and color cues to narrow down the detection regions to moving objects, people, and finally construction workers, respectively. The three cues are characterized by using background subtraction, the histogram of oriented gradients (HOG), and the HSV color histogram. The method has been tested on videos taken in various environments. The results demonstrate its suitability for automatic initialization of vision trackers.

Static and dynamic errors in particle tracking microrheology.

Particle tracking techniques are often used to assess the local mechanical properties of cells and biological fluids. The extracted trajectories are exploited to compute the mean-squared displacement that characterizes the dynamics of the probe particles. Limited spatial resolution and statistical uncertainty are the limiting factors that alter the accuracy of the mean-squared displacement estimation. We precisely quantified the effect of localization errors in the determination of the mean-squared displacement by separating the sources of these errors into two separate contributions. A "static error" arises in the position measurements of immobilized particles. A "dynamic error" comes from the particle motion during the finite exposure time that is required for visualization. We calculated the propagation of these errors on the mean-squared displacement. We examined the impact of our error analysis on theoretical model fluids used in biorheology. These theoretical predictions were verified for purely viscous fluids using simulations and a multiple-particle tracking technique performed with video microscopy. We showed that the static contribution can be confidently corrected in dynamics studies by using static experiments performed at a similar noise-to-signal ratio. This groundwork allowed us to achieve higher resolution in the mean-squared displacement, and thus to increase the accuracy of microrheology studies.

MoT - Mixture of trees probabilistic graphical model for video segmentation

We present a novel mixture of trees (MoT) graphical model for video segmentation. Each component in this mixture represents a tree structured temporal linkage between super-pixels from the first to the last frame of a video sequence. Our time-series model explicitly captures the uncertainty in temporal linkage between adjacent frames which improves segmentation accuracy. We provide a variational inference scheme for this model to estimate super-pixel labels and their confidences in nearly realtime. The efficacy of our approach is demonstrated via quantitative comparisons on the challenging SegTrack joint segmentation and tracking dataset [23].

Monitoring Activities from Multiple Video Streams: Establishing a Common Coordinate Frame

Passive monitoring of large sites typically requires coordination between multiple cameras, which in turn requires methods for automatically relating events between distributed cameras. This paper tackles the problem of self-calibration of multiple cameras which are very far apart, using feature correspondences to determine the camera geometry. The key problem is finding such correspondences. Since the camera geometry and photometric characteristics vary considerably between images, one cannot use brightness and/or proximity constraints. Instead we apply planar geometric constraints to moving objects in the scene in order to align the scene"s ground plane across multiple views. We do not assume synchronized cameras, and we show that enforcing geometric constraints enables us to align the tracking data in time. Once we have recovered the homography which aligns the planar structure in the scene, we can compute from the homography matrix the 3D position of the plane and the relative camera positions. This in turn enables us to recover a homography matrix which maps the images to an overhead view. We demonstrate this technique in two settings: a controlled lab setting where we test the effects of errors in internal camera calibration, and an uncontrolled, outdoor setting in which the full procedure is applied to external camera calibration and ground plane recovery. In spite of noise in the internal camera parameters and image data, the system successfully recovers both planar structure and relative camera positions in both settings.

Skin Color-Based Video Segmentation under Time-Varying Illumination

A novel approach for real-time skin segmentation in video sequences is described. The approach enables reliable skin segmentation despite wide variation in illumination during tracking. An explicit second order Markov model is used to predict evolution of the skin-color (HSV) histogram over time. Histograms are dynamically updated based on feedback from the current segmentation and predictions of the Markov model. The evolution of the skin-color distribution at each frame is parameterized by translation, scaling and rotation in color space. Consequent changes in geometric parameterization of the distribution are propagated by warping and resampling the histogram. The parameters of the discrete-time dynamic Markov model are estimated using Maximum Likelihood Estimation, and also evolve over time. The accuracy of the new dynamic skin color segmentation algorithm is compared to that obtained via a static color model. Segmentation accuracy is evaluated using labeled ground-truth video sequences taken from staged experiments and popular movies. An overall increase in segmentation accuracy of up to 24% is observed in 17 out of 21 test sequences. In all but one case the skin-color classification rates for our system were higher, with background classification rates comparable to those of the static segmentation.

Extraction and Clustering of Motion Trajectories in Video

A system is described that tracks moving objects in a video dataset so as to extract a representation of the objects' 3D trajectories. The system then finds hierarchical clusters of similar trajectories in the video dataset. Objects' motion trajectories are extracted via an EKF formulation that provides each object's 3D trajectory up to a constant factor. To increase accuracy when occlusions occur, multiple tracking hypotheses are followed. For trajectory-based clustering and retrieval, a modified version of edit distance, called longest common subsequence (LCSS) is employed. Similarities are computed between projections of trajectories on coordinate axes. Trajectories are grouped based, using an agglomerative clustering algorithm. To check the validity of the approach, experiments using real data were performed.

Head Tracking via Robust Registration in Texture Map Images

A novel method for 3D head tracking in the presence of large head rotations and facial expression changes is described. Tracking is formulated in terms of color image registration in the texture map of a 3D surface model. Model appearance is recursively updated via image mosaicking in the texture map as the head orientation varies. The resulting dynamic texture map provides a stabilized view of the face that can be used as input to many existing 2D techniques for face recognition, facial expressions analysis, lip reading, and eye tracking. Parameters are estimated via a robust minimization procedure; this provides robustness to occlusions, wrinkles, shadows, and specular highlights. The system was tested on a variety of sequences taken with low quality, uncalibrated video cameras. Experimental results are reported.

Active Hidden Models for Tracking with Kernel Projections

We introduce Active Hidden Models (AHM) that utilize kernel methods traditionally associated with classification. We use AHMs to track deformable objects in video sequences by leveraging kernel projections. We introduce the "subset projection" method which improves the efficiency of our tracking approach by a factor of ten. We successfully tested our method on facial tracking with extreme head movements (including full 180-degree head rotation), facial expressions, and deformable objects. Given a kernel and a set of training observations, we derive unbiased estimates of the accuracy of the AHM tracker. Kernels are generally used in classification methods to make training data linearly separable. We prove that the optimal (minimum variance) tracking kernels are those that make the training observations linearly dependent.

Improved Tracking of Multiple Humans with Trajectory Prediction and Occlusion Modeling

A combined 2D, 3D approach is presented that allows for robust tracking of moving bodies in a given environment as observed via a single, uncalibrated video camera. Tracking is robust even in the presence of occlusions. Low-level features are often insufficient for detection, segmentation, and tracking of non-rigid moving objects. Therefore, an improved mechanism is proposed that combines low-level (image processing) and mid-level (recursive trajectory estimation) information obtained during the tracking process. The resulting system can segment and maintain the tracking of moving objects before, during, and after occlusion. At each frame, the system also extracts a stabilized coordinate frame of the moving objects. This stabilized frame is used to resize and resample the moving blob so that it can be used as input to motion recognition modules. The approach enables robust tracking without constraining the system to know the shape of the objects being tracked beforehand; although, some assumptions are made about the characteristics of the shape of the objects, and how they evolve with time. Experiments in tracking moving people are described.

3D Trajectory Recovery for Tracking Multiple Objects and Trajectory Guided Recognition of Actions

A mechanism is proposed that integrates low-level (image processing), mid-level (recursive 3D trajectory estimation), and high-level (action recognition) processes. It is assumed that the system observes multiple moving objects via a single, uncalibrated video camera. A novel extended Kalman filter formulation is used in estimating the relative 3D motion trajectories up to a scale factor. The recursive estimation process provides a prediction and error measure that is exploited in higher-level stages of action recognition. Conversely, higher-level mechanisms provide feedback that allows the system to reliably segment and maintain the tracking of moving objects before, during, and after occlusion. The 3D trajectory, occlusion, and segmentation information are utilized in extracting stabilized views of the moving object. Trajectory-guided recognition (TGR) is proposed as a new and efficient method for adaptive classification of action. The TGR approach is demonstrated using "motion history images" that are then recognized via a mixture of Gaussian classifier. The system was tested in recognizing various dynamic human outdoor activities; e.g., running, walking, roller blading, and cycling. Experiments with synthetic data sets are used to evaluate stability of the trajectory estimator with respect to noise.

Trajectory Guided Tracking and Recognition of Actions

A combined 2D, 3D approach is presented that allows for robust tracking of moving people and recognition of actions. It is assumed that the system observes multiple moving objects via a single, uncalibrated video camera. Low-level features are often insufficient for detection, segmentation, and tracking of non-rigid moving objects. Therefore, an improved mechanism is proposed that integrates low-level (image processing), mid-level (recursive 3D trajectory estimation), and high-level (action recognition) processes. A novel extended Kalman filter formulation is used in estimating the relative 3D motion trajectories up to a scale factor. The recursive estimation process provides a prediction and error measure that is exploited in higher-level stages of action recognition. Conversely, higher-level mechanisms provide feedback that allows the system to reliably segment and maintain the tracking of moving objects before, during, and after occlusion. The 3D trajectory, occlusion, and segmentation information are utilized in extracting stabilized views of the moving object that are then used as input to action recognition modules. Trajectory-guided recognition (TGR) is proposed as a new and efficient method for adaptive classification of action. The TGR approach is demonstrated using "motion history images" that are then recognized via a mixture-of-Gaussians classifier. The system was tested in recognizing various dynamic human outdoor activities: running, walking, roller blading, and cycling. Experiments with real and synthetic data sets are used to evaluate stability of the trajectory estimator with respect to noise.

Inferring Body Pose without Tracking Body Parts

A novel approach for estimating articulated body posture and motion from monocular video sequences is proposed. Human pose is defined as the instantaneous two dimensional configuration (i.e., the projection onto the image plane) of a single articulated body in terms of the position of a predetermined set of joints. First, statistical segmentation of the human bodies from the background is performed and low-level visual features are found given the segmented body shape. The goal is to be able to map these, generally low level, visual features to body configurations. The system estimates different mappings, each one with a specific cluster in the visual feature space. Given a set of body motion sequences for training, unsupervised clustering is obtained via the Expectation Maximation algorithm. Then, for each of the clusters, a function is estimated to build the mapping between low-level features to 3D pose. Currently this mapping is modeled by a neural network. Given new visual features, a mapping from each cluster is performed to yield a set of possible poses. From this set, the system selects the most likely pose given the learned probability distribution and the visual feature similarity between hypothesis and input. Performance of the proposed approach is characterized using a new set of known body postures, showing promising results.