Color based moving object tracking with an active camera using motion information

Çelik, Yunus; Altun, Mahmut; Güneş, Mahit

doi:10.1109/idap.2017.8090332

Cited by 13 publications

(6 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There exists not much research attention in the area of active tracking. Conventional solutions dealt with object tracking and camera control in separate components [23], [24], [25], [26], [27], [28]. For example, in [27], object detection is applied to estimate motion and a camera with pan and tilt operations is employed to keep the object in the field of view.…”

Section: Object Trackingmentioning

confidence: 99%

End-to-End Active Object Tracking and Its Real-World Deployment via Reinforcement Learning

Luo

Sun

Zhong

et al. 2020

IEEE Trans. Pattern Anal. Mach. Intell.

128

View full text Add to dashboard Cite

We study active object tracking, where a tracker takes visual observations (i.e., frame sequences) as input and produces the corresponding camera control signals as output (e.g., move forward, turn left, etc.). Conventional methods tackle tracking and camera control tasks separately, and the resulting system is difficult to tune jointly. These methods also require significant human efforts for image labeling and expensive trial-and-error system tuning in the real world. To address these issues, we propose, in this paper, an end-to-end solution via deep reinforcement learning. A ConvNet-LSTM function approximator is adopted for the direct frame-to-action prediction. We further propose an environment augmentation technique and a customized reward function, which are crucial for successful training. The tracker trained in simulators (ViZDoom and Unreal Engine) demonstrates good generalization behaviors in the case of unseen object moving paths, unseen object appearances, unseen backgrounds, and distracting objects. The system is robust and can restore tracking after occasional lost of the target being tracked. We also find that the tracking ability, obtained solely from simulators, can potentially transfer to real-world scenarios. We demonstrate successful examples of such transfer, via experiments over the VOT dataset and the deployment of a real-world robot using the proposed active tracker trained in simulation.

show abstract

Section: Object Trackingmentioning

confidence: 99%

End-to-End Active Object Tracking and Its Real-World Deployment via Reinforcement Learning

Luo

Sun

Zhong

et al. 2020

IEEE Trans. Pattern Anal. Mach. Intell.

128

View full text Add to dashboard Cite

show abstract

“…The principle segments utilized for setting up the mechanical vehicle are appeared in Figure 6. These parts are Raspberry Pi III model B, the L298N motor driver, the ultrasonic sensor, the voltage steps down regulator, the Li-Po battery, and the RPi camera [20][21][22][23][24][25]. Figure 6.…”

Section: Car Parts Gathering and Arrangementmentioning

confidence: 99%

Software engineering model based smart indoor localization system using deep-learning

Resan

Croock

2020

TELKOMNIKA

View full text Add to dashboard Cite

During the last few years, the allocation of objects or persons inside a specific building is highly required. It is well known that the global positioning system (GPS) cannot be adopted in indoor environment due to the lack of signals. Therefore, it is important to discover a new way that works inside. The proposed system uses the deep learning techniques to classify places based on capturing images. The proposed system contains two parts: software part and hardware part. The software part is built based on software engineering model to increase the reliability, flexibility, and scalability. In addition, this part, the dataset is collected using the Raspberry Pi III camera as training and validating data set. This dataset is used as an input to the proposed deep learning model. In the hardware part, Raspberry Pi III is used for loading the proposed model and producing prediction results and a camera that is used to collect the images dataset. Two wheels' car is adopted as an object for introducing indoor localization project. The obtained training accuracy is 99.6% for training dataset and 100% for validating dataset.

show abstract

“…In [28], a hierarchical system to control a set of surveillance cameras is presented. In [5], the authors propose a system composed by three main different components to perform mapping, detecting and tracking the object. [29] introduces a new modular architecture that incorporates a model for perception and another one for the control policy.…”

Section: Related Workmentioning

confidence: 99%

“…Therefore, in this work, we focus on the more challenging task of Visual Active Tracking (VAT) [5], [6], [7], in which the tracker has to actively search for and track a specified target (Fig. 1).…”

Section: Introductionmentioning

confidence: 99%

Enhancing Continuous Control of Mobile Robots for End-to-End Visual Active Tracking

Devo¹,

Dionigi²

2020

Preprint

View full text Add to dashboard Cite

In the last decades, visual target tracking has been one of the primary research interests of the Robotics research community. The recent advances in Deep Learning technologies have made the exploitation of visual tracking approaches effective and possible in a wide variety of applications, ranging from automotive to surveillance and human assistance. However, the majority of the existing works focus exclusively on passive visual tracking, i.e., tracking elements in sequences of images by assuming that no actions can be taken to adapt the camera position to the motion of the tracked entity. On the contrary, in this work, we address visual active tracking, in which the tracker has to actively search for and track a specified target. Current State-of-the-Art approaches use Deep Reinforcement Learning (DRL) techniques to address the problem in an endto-end manner. However, two main problems arise: i) most of the contributions focus only on discrete action spaces and the ones that consider continuous control do not achieve the same level of performance; and ii) if not properly tuned, DRL models can be challenging to train, resulting in a considerably slow learning progress and poor final performance. To address these challenges, we propose a novel DRL-based visual active tracking system that provides continuous action policies. To accelerate training and improve the overall performance, we introduce additional objective functions and a Heuristic Trajectory Generator (HTG) to facilitate learning. Through an extensive experimentation, we show that our method can reach and surpass other State-of-the-Art approaches performances, and demonstrate that, even if trained exclusively in simulation, it can successfully perform visual active tracking even in real scenarios.

show abstract

Color based moving object tracking with an active camera using motion information

Cited by 13 publications

References 5 publications

End-to-End Active Object Tracking and Its Real-World Deployment via Reinforcement Learning

End-to-End Active Object Tracking and Its Real-World Deployment via Reinforcement Learning

Software engineering model based smart indoor localization system using deep-learning

Enhancing Continuous Control of Mobile Robots for End-to-End Visual Active Tracking

Contact Info

Product

Resources

About