Tracking Stage Research Articles

Barriers to Full Participation in the Open Science Life Cycle among Early Career Researchers

Open science (OS) is currently dominated by a small subset of practices that occur late in the scientific process. Early career researchers (ECRs) will play a key role in transitioning the scientific community to more widespread use of OS from pre-registration to publication, but they also face unique challenges in adopting these practices. Here, we discuss these challenges across the OS life cycle. Our essay relies on the published literature, an informal survey of 32 ECRs from 14 countries, and discussions among members of the Global Working Group on Open Science (Global Young Academy and National Young Academies). We break the OS life cycle into four stages—study design and tracking (pre-registration, open processes), data collection (citizen science, open hardware, open software, open data), publication (open access publishing, open peer review, open data), and outreach (open educational resources, citizen science)—and map potential barriers at each stage. The most frequently discussed barriers across the OS life cycle were a lack of awareness and training, prohibitively high time commitments, and restrictions and/or a lack of incentives by supervisors. We found that OS practices are highly fragmented and that awareness is particularly low for OS practices that occur during the study design and tracking stage, possibly creating ‘path-dependencies’ that reduce the likelihood of OS practices at later stages. We note that, because ECRs face unique barriers to adopting OS, there is a need for specifically targeted policies such as mandatory training at the graduate level and promotion incentives.

Thermal infrared object tracking via unsupervised deep correlation filters

In thermal infrared (TIR) object tracking, the scarcity of annotated TIR image datasets hampers the performance of deep learning-based trackers. Therefore, most trackers still only use handcrafted features. To tackle this problem, we propose an unsupervised deep correlation TIR object tracking framework. Motivated by the forward-backward tracking consistency of robust trackers, we explore a novel approach that learns a lightweight feature extraction network for TIR images in an unsupervised manner. We follow a Siamese correlation filter framework to construct the network and train it based on sample quality-sensitive consistency loss. To the best of our knowledge, this is the first unsupervised end-to-end trained feature extraction network for TIR object tracking. At the tracking stage, to improve tracking performance, we utilize continuous multichannel correlation filters to combine CNN features with handcrafted features. Extensive experiments indicate that the proposed method outperforms state-of-the-art trackers with more than 10% relative gain in EAO, clearly showing its effectiveness and robustness.

High-Performance Visual Tracking Based on High-Order Pooling Network

Convolution Neural Network (CNN) features have been widely used in visual tracking due to their powerful representation. As an important component of CNN, the pooling layer plays a critical role, but the max/average/min operation only explores the first-order information, which limits the discrimination ability of the CNN features in some complex situations. In this paper, a high-order pooling layer is integrated into the VGG16 network for visual tracking. In detail, a high-order covariance pooling layer is employed to replace the last maxpooling layer to learn discrimination features and is trained on the ImageNet and CUB200-2011 data sets. In tracking stage, the multiple levels of feature maps are extracted as the appearance representation of the target. After that, the extracted CNN features are integrated into the correlation filters framework when tracking is on-the-fly. The experimental results show that the proposed algorithm achieves excellent performance in both success rate and tracking accuracy.

Moving Object Detection over Wireless Visual Sensor Networks using Spectral Dual Mode Background Subtraction

Wireless Visual Sensor Networks (WVSN) play an essential role in tracking moving objects. WVSN's key drawbacks are storage, power, and bandwidth. Background subtraction is used in the early stages of target tracking to extract moving targets from video images. Many standard methods of subtracting backgrounds are no longer suitable for embedded devices because they use complex statistical models to manage small changes in lighting. This paper introduces a system based on the Partial Discrete Cosine Transform (PDCT), reducing the vast dimensions of processed data while retaining most of the important information, thereby reducing processing and transmission energy. It also uses a dual-mode single Gaussian model (SGM) for accurate detection of moving objects. The proposed system's performance is to be assessed using the standard CDnet 2014 benchmark dataset in terms of detection accuracy and time complexity. Furthermore, the suggested method is compared to previous WVSN background subtraction methods. Simulation results show that the proposed method consistently has 15% better accuracy and is up to 3 times faster than the state-of-the-art object detection methods for WVSN. Finally, we showed the practicality of the suggested method by simulating it in a sensor network environment using the Contiki OS Cooja Simulator and implementing it in a real testbed using Cortex M3 open nodes of IOT-LAB.

Robust Correlation Tracking in Unmanned Aerial Vehicle Videos via Deep Target-Specific Rectification Networks

Unmanned Aerial Vehicle (UAV) Videos have received active research attention in the remote sensing field by taking full advantage of the bird’s eye view offered by UAV. At the same time, visual tracking approaches based on discriminative correlation filters (DCF) have recently achieved increasing popularity and success. Despite their success, the tracking robustness and accuracy of existing DCF-based trackers are hard to promote in challenging tracking scenarios of aerial videos due to excessive reliance on the response map and fixed linear model update strategy. To resolve these issues, we propose a robust DCF-based tracking framework via an effective pretrained rectification network for UAV-based remote sensing. Specifically, the target-specific rectification network is offline trained to discriminatively classify the target and background. During the online tracking stage, the DCF module performs fast inference to obtain the potential locations of the target. After that, the deep rectification network evaluates the correlation-specific proposals offered by the DCF module and provides precise tracking results. Besides, to achieve a robust and adaptive model update strategy, we propose to finetune both the DCF module and rectification network according to the classification confidence of the estimated result. Extensive experimental results on recent UAV benchmarks demonstrate that our method achieves better performance than other competing algorithms.

Multilayer Perceptron Generative Model via Adversarial Learning for Robust Visual Tracking

Visual tracking is an open and exciting field of research. The researchers introduced great efforts to be close to the ideal state of stable tracking of objects regardless of different appearances or circumstances. Owing to the attractive advantages of generative adversarial networks (GANs), they have been a promising area of research in many fields. However, GAN network architecture has not been thoroughly investigated in the visual tracking research community. Inspired by visual tracking via adversarial learning (VITAL), we present a novel network to generate randomly initialized masks for building augmented feature maps using multilayer perceptron (MLP) generative models. To obtain more robust tracking these augmented masks can extract robust features that do not change over a long temporal span. Some models such as deep convolutional generative adversarial networks (DCGANs) have been proposed to obtain powerful generator architectures by eliminating or minimizing the use of fully connected layers. This study demonstrates that the use of MLP architecture for the generator is more robust and efficient than the convolution-only architecture. Also, to realize better performance, we used one-sided label smoothing to regularize the discriminator in the training stage and the label smoothing regularization (LSR) method to reduce the overfitting of the classifier in the online tracking stage. The experiments show that the proposed model is more robust than the DCGAN model and offers satisfactory performance compared with the state-of-the-art deep visual trackers on OTB-100, VOT2019 and LaSOT datasets.

Design of a Fuzzy Fractional Order Adaptive Impedance Controller with Integer Order Approximation for Stable Robotic Contact Force Tracking in Uncertain Environment

Abstract Current research in robot compliance control is unable to take both transient contact force overshoots and steady-state force tracking problems into account. To address this problem, we propose a fuzzy fractional order (FO) adaptive impedance controller to avoid the force overshoots in the contact stage while keeping force error in the dynamic tracking stage, where traditional control algorithms are not competent. A percentage gain is adopted to map FO parameters to integer order (IO) parameters by their natural properties, and a fuzzy logical controller is introduced to improve the system stability. The simulation results indicate that the proposed controller can be made more stable than and superior to the general impedance controller, and the force tracking results also have been compared with the previous control methods.

A New Method for Surface-to-Air Video Detection and Tracking of Airborne Vehicles

Detection process of airborne targets may be thought simple because of the incompatible nature of aircraft, choppers, UAVs, and drones regarding clear sky background. When changes in the background are considered, brightness variation of the sky complicates the process. Changes in the shapes and types of clouds add another challenge to the process. Tracking process directly depends on the detection process and type of the data stream. The practical systems used for video detection and tracking of airborne targets are manual, and manual structures have some drawbacks compared to automatic structures. For video surveillance, guidance, regional security, and defense applications in dense environments, automatic detection and tracking process may be an obligation rather than preference. In this study, an automatic detection and tracking algorithm for video streams of airborne targets is proposed. A land-based moving camera captures the video data, and not only the flying objects but probably also the camera are in motion. Although the detection and tracking of moving objects via moving sensors is a relatively arduous task, this is the prevalent case in real-life scenarios. Video detection and tracking systems have one or more moving video sensors, while one or more flying air vehicles are in operation area. The proposed algorithm includes an image processing stage for detection and a tracking stage for initiation and continuation. An assessment study has been conducted for the actual video data and found that the proposed method yields successful results for detection, track formation, and continuation processes.

Long Distance Moving Vehicle Tracking with a Multirotor Based on IMM-Directional Track Association

The multirotor has the capability to capture distant objects. Because the computing resources of the multirotor are limited, efficiency is an important factor to consider. In this paper, multiple target tracking with a multirotor at a long distance (~400 m) is addressed; the interacting multiple model (IMM) estimator combined with the directional track-to-track association (abbreviated as track association) is proposed. The previous work of the Kalman estimator with the track association approach is extended to the IMM estimator with the directional track association. The IMM estimator can handle multiple targets with various maneuvers. The track association scheme is modified in consideration of the direction of the target movement. The overall system is composed of moving object detection for measurement generation and multiple target tracking for state estimation. The moving object detection consists of frame-to-frame subtraction of three-color layers and thresholding, morphological operation, and false alarm removing based on the object size and shape properties. The centroid of the detected object is input into the next tracking stage. The track is initialized using the difference between two nearest points measured in consecutive frames. The measurement nearest to the state prediction is used to update the state of the target for measurement-to-track association. The directional track association tests both the hypothesis and the maximum deviation between the displacement and directions of two tracks followed by track selection, fusion, and termination. In the experiment, a multirotor flying at an altitude of 400 m captured 55 moving vehicles around a highway interchange for about 20 s. The tracking performance is evaluated for the IMMs using constant velocity (CV) and constant acceleration (CA) motion models. The IMM-CA with the directional track association scheme outperforms other methods with an average total track life of 91.7% and an average mean track life of 84.2%.

Real‐time traffic anomaly detection based on Gaussian mixture model and hidden Markov model

AbstractWith the rapid development and wide application of visual sensor technology and image processing technology, the traffic anomaly events detection methods of intelligent transportation system (ITS) are being constantly updated. Traditional machine learning methods do not require high computing power but lack high accuracy, while deep learning methods provide high accuracy but demand of high computing power for edge computing hardware. To improve accuracy and real‐time performance of detecting abnormal traffic events in edge computing hardware, a novel traffic anomaly detection model was developed in the following steps. In the first moving object tracking stage, a hybrid model was designed by combining Gaussian mixture model (GMM) with hidden Markov model, and optimized the tracking accuracy of target trajectories in multiframe images from traffic surveillance video. In the second abnormal events classification stage, principal component analysis was applied to reduce the dimension of the trajectories features, and then the abnormal traffic behaviors were classified by K‐Nearest Neighbor. Experiments demonstrate that the model can achieve higher detection accuracy than GMM method as well as a faster detection speed than You only look once deep learning method in moving object tracking stage.

Charged Particle Tracking via Edge-Classifying Interaction Networks

Recent work has demonstrated that geometric deep learning methods such as graph neural networks (GNNs) are well suited to address a variety of reconstruction problems in high-energy particle physics. In particular, particle tracking data are naturally represented as a graph by identifying silicon tracker hits as nodes and particle trajectories as edges, given a set of hypothesized edges, edge-classifying GNNs identify those corresponding to real particle trajectories. In this work, we adapt the physics-motivated interaction network (IN) GNN toward the problem of particle tracking in pileup conditions similar to those expected at the high-luminosity Large Hadron Collider. Assuming idealized hit filtering at various particle momenta thresholds, we demonstrate the IN’s excellent edge-classification accuracy and tracking efficiency through a suite of measurements at each stage of GNN-based tracking: graph construction, edge classification, and track building. The proposed IN architecture is substantially smaller than previously studied GNN tracking architectures; this is particularly promising as a reduction in size is critical for enabling GNN-based tracking in constrained computing environments. Furthermore, the IN may be represented as either a set of explicit matrix operations or a message passing GNN. Efforts are underway to accelerate each representation via heterogeneous computing resources towards both high-level and low-latency triggering applications.

Система мультикамерного отслеживания объекта

The article proposes a multi-camera tracking system for an object, implemented using computer vision technologies and allowing the video surveillance operator in real time to select an object that will be monitored by the system in future. It will be ready to give out the location of the object at any time. The solution to this problem is divided into three main stages: the detection stage, the tracking stage and the stage of interaction of several cameras. Methods of detection, tracking of objects and the interaction of several cameras have been investigated. To solve the problem of detection, the method of optical flow and the method of removing the background were investigated, to solve the problem of tracking — the method of matching key points and the correlation method, to solve the problem of interaction between several surveillance cameras — the method of the topological graph of a network of cameras. An approach is proposed for constructing a system that uses a combination of the background removal method, the correlation method and the method of the topological graph of a network of cameras. The stages of detection and tracking have been experimentally implemented, that is, the task of tracking an object within the coverage area of one video camera has been solved. The implemented system showed good results: a sufficiently high speed and accuracy with rare losses of the tracked object and with a slight decrease in the frame rate.

Efficient detection and robust tracking of spermatozoa in microscopic video

Sperm concentration and motility are generally analysed only in the discrete state in microscopic videos. As for sperm nonspecific aggregation areas, it brings difficulties to accurate sperm detection. In this paper, an algorithm for nonspecific aggregates automatic segmentation, detection and tracking of sperm is proposed. A grid model commensurate with the size of a sperm head is created to segment nonspecific aggregation areas. Multi-scale edge function and new energy functional are designed based on the level set method to realize sperm head segmentation. In the sperm tracking stage, we improve the weight condition and the standard of trust flow quantization based on graph theory method, and simplify the sperm tracking to the vertex matching between two frames to solve the matching failure problem of adjacent frames with small space distance. The proposed method achieves accurate segmentation of sperm non-specific aggregation regions, which outperforms the level set methods of LBF and SBGFR. At the same time, our method can real-time calculation of sperm concentration and motility during sperm tracking. It is compared with four state-of-the-art algorithms, and it gives lower tracking error rates, which has potential applications in male fertility field.

A Coarse-to-Fine Object Tracking Based on Attention Mechanism

In the visual object tracking task, although the convolutional neural network has excellent performance, it is difficult to learn the global and long-range semantic information interaction due to the inherent locality of convolution. Simultaneously, the existing template updating methods focus on the target position of the next frame, which will integrate more irrelevant background and cause tracking instability. Therefore, to address this issue, inspired by Transformer, in this paper, we propose a coarse-to-fine tracking method based on the attention mechanism (CTFT). First, employing swin transformer block instead of the original convolutional network backbone to achieve self-attention from local to global. Second, using attention to effectively combine template and search region features in the part of feature fusion, and explore the possibility of using attention mechanism as the backbone in object tracking for the first time. Finally, a tracking strategy from coarse to fine is proposed. In the offline coarse tracking stage, the initial estimation of the target object is generated to obtain the coarse regress bounding box of the target. In the online fine tracking stage, we make use of the coarse regress bounding box to expand the corresponding object frame, and use the segmentation to get the fine regress bounding box of the object. Experiments on challenging benchmarks including GOT-10k, LaSOT, VOT2018, VOT2019, UAV123 and OTB-100 demonstrate that the proposed CTFT outperforms many state-of-the-art trackers and achieves leading performance.

Detection and Tracking of Pedestrians Using Doppler LiDAR

Pedestrian detection and tracking is necessary for autonomous vehicles and traffic management. This paper presents a novel solution to pedestrian detection and tracking for urban scenarios based on Doppler LiDAR that records both the position and velocity of the targets. The workflow consists of two stages. In the detection stage, the input point cloud is first segmented to form clusters, frame by frame. A subsequent multiple pedestrian separation process is introduced to further segment pedestrians close to each other. While a simple speed classifier is capable of extracting most of the moving pedestrians, a supervised machine learning-based classifier is adopted to detect pedestrians with insignificant radial velocity. In the tracking stage, the pedestrian’s state is estimated by a Kalman filter, which uses the speed information to estimate the pedestrian’s dynamics. Based on the similarity between the predicted and detected states of pedestrians, a greedy algorithm is adopted to associate the trajectories with the detection results. The presented detection and tracking methods are tested on two data sets collected in San Francisco, California by a mobile Doppler LiDAR system. The results of the pedestrian detection demonstrate that the proposed two-step classifier can improve the detection performance, particularly for detecting pedestrians far from the sensor. For both data sets, the use of Doppler speed information improves the F1-score and the recall by 15% to 20%. The subsequent tracking from the Kalman filter can achieve 83.9–55.3% for the multiple object tracking accuracy (MOTA), where the contribution of the speed measurements is secondary and insignificant.

Terahertz Ultra-Massive MIMO-Based Aeronautical Communications in Space-Air-Ground Integrated Networks

The emerging space-air-ground integrated network has attracted intensive research and necessitates reliable and efficient aeronautical communications. This paper investigates terahertz Ultra-Massive (UM)-MIMO-based aeronautical communications and proposes an effective channel estimation and tracking scheme, which can solve the performance degradation problem caused by the unique triple delay-beam-Doppler squint effects of aeronautical terahertz UM-MIMO channels. Specifically, based on the rough angle estimates acquired from navigation information, an initial aeronautical link is established, where the delay-beam squint at transceiver can be significantly mitigated by employing a Grouping True-Time Delay Unit (GTTDU) module (e.g., the designed Rotman lens -based GTTDU module). According to the proposed prior-aided iterative angle estimation algorithm, azimuth/elevation angles can be estimated, and these angles are adopted to achieve precise beam-alignment and refine GTTDU module for further eliminating delay-beam squint. Doppler shifts can be subsequently estimated using the proposed prior-aided iterative Doppler shift estimation algorithm. On this basis, path delays and channel gains can be estimated accurately, where the Doppler squint can be effectively attenuated via compensation process. For data transmission, a data-aided decision-directed based channel tracking algorithm is developed to track the beam-aligned effective channels. When the data-aided channel tracking is invalid, angles will be re-estimated at the pilot-aided channel tracking stage with an equivalent sparse digital array, where angle ambiguity can be resolved based on the previously estimated angles. The simulation results and the derived Cramer-Rao lower bounds verify the effectiveness of our solution.

A 56 Gbps 4‐tap PAM‐4 direct decision feedback equaliser with negative capacitance employing dynamic CML comparators in 65‐nm CMOS

Here, a 4-level pulse amplitude modulation direct decision feedback equaliser (DFE) with a novel dynamic current-mode-logic comparator (DCMLC) is presented. The DCMLC breaks the trade-off between settling time and regeneration time in traditional CML comparator design by utilizing dynamic logic and separately optimizes the tracking stage and regeneration stage for a correct latch operation at ultrahigh speed. Compared with the traditional CML comparator, the DCMLC reduces delay by 36% and has better input sensitivity on high baud rates at the cost of 7% shrunk output swing. The negative capacitance is adopted to achieve a 0.5 dB bandwidth extension ratio of up to 1.89. The reduced delay and wider bandwidth of the proposed comparator allow the implementation of 4-tap direct DFE at 56 Gbps with 2.8 pJ/bit energy efficiency and an active area of 0.007 mm2 in 65-nm CMOS technology.

Motion Estimation Using Region-Level Segmentation and Extended Kalman Filter for Autonomous Driving

Motion estimation is crucial to predict where other traffic participants will be at a certain period of time, and accordingly plan the route of the ego-vehicle. This paper presents a novel approach to estimate the motion state by using region-level instance segmentation and extended Kalman filter (EKF). Motion estimation involves three stages of object detection, tracking and parameter estimate. We first use a region-level segmentation to accurately locate the object region for the latter two stages. The region-level segmentation combines color, temporal (optical flow), and spatial (depth) information as the basis for segmentation by using super-pixels and Conditional Random Field. The optical flow is then employed to track the feature points within the object area. In the stage of parameter estimate, we develop a relative motion model of the ego-vehicle and the object, and accordingly establish an EKF model for point tracking and parameter estimate. The EKF model integrates the ego-motion, optical flow, and disparity to generate optimized motion parameters. During tracking and parameter estimate, we apply edge point constraint and consistency constraint to eliminate outliers of tracking points so that the feature points used for tracking are ensured within the object body and the parameter estimates are refined by inner points. Experiments have been conducted on the KITTI dataset, and the results demonstrate that our method presents excellent performance and outperforms the other state-of-the-art methods either in object segmentation and parameter estimate.

Moving Object Detection Technology of Line Dancing Based on Machine Vision

In this paper, line dancing's moving object detection technology based on machine vision is studied to improve object detection. For this purpose, the improved frame difference for the background modeling technique is combined with the target detection algorithm. The moving target is extracted, and the postmorphological processing is carried out to make the target detection more accurate. Based on this, the tracking target is determined on the time axis of the moving target tracking stage, the position of the target in each frame is found, and the most similar target is found in each frame of the video sequence. The association relationship is established to determine a moving object template or feature. Through certain measurement criteria, the mean-shift algorithm is used to search the optimal candidate target in the image frame and carry out the corresponding matching to realize moving objects' tracking. This method can detect the moving targets of line dancing in various areas through the experimental analysis, which will not be affected by the position or distance, and always has a more accurate detection effect.

Remote sensing target tracking in satellite videos based on a variable‐angle‐adaptive Siamese network

Remote sensing target tracking in satellite videos plays a key role in various fields. However, due to the complex backgrounds of satellite video sequences and many rotation changes of highly dynamic targets, typical target tracking methods for natural scenes cannot be used directly for such tasks, and their robustness and accuracy are difficult to guarantee. To address these problems, an algorithm is proposed for remote sensing target tracking in satellite videos based on a variable-angle-adaptive Siamese network (VAASN). Specifically, the method is based on the fully convolutional Siamese network (Siamese-FC). First, for the feature extraction stage, to reduce the impact of complex backgrounds, we present a new multifrequency feature representation method and introduce the octave convolution (OctConv) into the AlexNet architecture to adapt to the new feature representation. Then, for the tracking stage, to adapt to changes in target rotation, a variable-angle-adaptive module that uses a fast text detector with a single deep neural network (TextBoxes++) is introduced to extract angle information from the template frame and detection frames and performs angle consistency update operations on the detection frames. Finally, qualitative and quantitative experiments using satellite datasets show that the proposed method can improve tracking accuracy while achieving high efficiency.