Human Tracking Applications Research Articles

Evolution of CNNs in human tracking applications

Tracking the movement of objects in videos or footage from CCTV systems plays an integral role in crime investigations, surveillance, security predictions, and many other domains. Historically, this task was primarily entrusted to dedicated observers or analysts who would be summoned to review pre-recorded footage post-event. With the advent of machine learning and AI, convolutional neural networks (CNNs) have paved the way for computers to augment human capabilities in analyzing streaming videos or archived recordings. Among the various tracking methodologies that machine learning offers, YOLO (You Only Look Once) and R-CNN (Region-based Convolutional Neural Networks), along with its iterations, stand out as some of the most reliable and precise. However, the scope of analysis often extends beyond these techniques. To enhance accuracy and provide an adept classification mechanism, the Deep SORT (Simple Online and Realtime Tracking) algorithm emerges as pivotal. Its synergy with human detection remains a significant area of discussion and will be deliberated upon in this study. This review aims to elucidate the intricacies of these state-of-the-art methods and their interplay in modern tracking systems.

Low-Cost Indoor Human Tracking by Utilizing Fluctuation of Received Radio Signal Strength

Existing indoor positioning technologies aim at providing location information of extremely high accuracy, but are limited by the requirement for infrastructure with high installation and maintenance costs. Thus, for indoor human tracking applications, a suitable indoor positioning technique should be selected by taking into account practical considerations of adequate accuracy as well as cost-effectiveness. Based on this observation, an indoor, room-level tracking method is presented, which is characterized by an infrastructure that is cost-effective to build and maintain with no need to deal with calibration issues of the fingerprint map caused by hardware heterogeneity. A mobile device tracks variations in the radio signal strength from nearby Bluetooth beacons to determine the entry and exit of the user’s device from a given room or hallway. Therefore, by determining the sequence of users’ movements, their moving path can be recorded for indoor tracking or navigation. Experiments revealed that the proposed method attains an average hit rate, false alarm rate, and miss rate of 94.2%, 3.6%, and 5.8%, respectively.

METHOD FOR INDOOR HUMAN POSITION TRACKING USING MULTIPLE DEPTH SENSORS

In this article a positioning method for covering room area was proposed. Multiple Kinect depth sensors were used to work around narrow field of view of one Kinect sensor and cover the room area to prevent blind spots. Affine transformation was used to convert coordinates of the Kinect sensors to the coordinates of the room. Indoor human tracking application was developed in research. During application testing, average aggregated error of 15cm was determined.

Development of Location Detection and Human Tracking Application via GPS and GSM Service: A Pilot Study

Development of Location Detection and Human Tracking Application via GPS and GSM Service: A Pilot Study

Toward Accurate Human Tracking: Modeling Time-of-Arrival for Wireless Wearable Sensors in Multipath Environment

For time-of-arrival-(TOA)-based indoor human tracking system, the wireless channel between human body surface and external reference node can be regarded as the source of inaccuracy. Since only the arrival time of direct path provides accurate range estimate, the nonline of sight caused by human body leads to undetectable direct path condition and thus results in a significant distance measurement error. In this paper, we measured TOA ranging error for indoor human tracking applications inside a typical office environment. A large number of TOA ranging samples was obtained and statistically analyzed. The TOA ranging error was modeled as a Gaussian random variable with the parameters, including position of target sensor, angle between human facing direction, and direction of transmitter–receiver, signal-to-noise ratio, and bandwidth of the system. As a validation of proposed model, excellent agreement has been found between empirical measurement and model-based software simulation.

Securing a wireless sensor network for human tracking: a review of solutions

SUMMARYCurrently, wireless sensor networks (WSNs) are formed by devices with limited resources and limited power energy availability. Thanks to their cost effectiveness, flexibility, and ease of deployment, wireless sensor networks have been applied to many scenarios such as industrial, civil, and military applications. For many applications, security is a primary issue, but this produces an extra energy cost. Thus, in real applications, a trade‐off is required between the security level and energy consumption. This paper evaluates different security schemes applied to human tracking applications, based on a real‐case scenario. Copyright © 2013 John Wiley & Sons, Ltd.

Adaptive Fuzzy Particle Filter Tracker for a PTZ Camera in an IP Surveillance System

We propose an adaptive fuzzy particle filter (PF) (AFPF) method adapted to general object tracking with an IP pan-tilt-zoom (PTZ) camera. PF samples are weighted using fuzzy membership functions and are applied to geometric and appearance features. In our PF, targets are modeled and tracked based on sampling around predicted positions obtained by a position predictor and moving regions detected by optical flow. Sample features are scored based on fuzzy rules. In this paper, we apply the AFPF to a human-tracking application in an IP PTZ surveillance system. Results show that our system has good target-detection precision (>; 93.9%), low track fragmentation, and a high processing rate, and the target is almost always located within one-sixth of the image diameter from the image center.

Control of Head-Eye Coordination of Conversational Robotic Actors

Head-eye coordination for conversational robotic actors require sophisticated control schemes in order to achieve realistic motion distribution between the head and the eyes. In turn, this requires a thorough understanding of human kinesiology that governs the way humans coordinate head-eye motion and the ability to control the motion of humanoid robotic actors to actuate in the same manner that humans do. In this paper, we propose an efficient head-eye motion coordination scheme to generate realistic head-eye movement of the conversational robotic actors for human tracking application using a combination of cost function optimization approach and image-based visual feedback. A cost function is formed based on kinesiology and then optimized to obtain a realistic head-eye trajectory. To minimize the effect from modelling error and imperfections of the actual humanoid hardware, visual feedback is implemented to ensure head-eye motion tracking accuracy. Experimental results show the tracking efficiency and realism of the motion generated by the proposed scheme with Lilly, a humanoid under development in our lab.