Tracking of Moving Objects With Regeneration of Object Feature Points

Lychkov, Igor I.; Alfimtsev, Alexander; Sakulin, Sergey

doi:10.1109/glosic.2018.8570061

Cited by 13 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The system is tested against objects of two categories namely humans and vehicles in motion. Lychkov et.al [4] proposed a novel object tracking system that addresses the issue of losing track of object's feature points. The authors use the concept of regeneration in living beings to detection of objects in video.…”

Section: Related Workmentioning

confidence: 99%

Object Recognition using Lucas-Kanade Technique and Support Vector Machine Based Classification in Video Surveillance Systems

Pillai,

2019

IJEAT

View full text Add to dashboard Cite

Object recognition in video surveillance systems is the primary and most significant challenge task in the field of image processing. Video Surveillance systems provides us continuous monitoring of the objects for the enhancement of security and control. This paper presents novel approach recognizing the objects using Shi-Tomasi approach for detecting the corners of the object and then applies the Lucas-Kanade techniques to extract the features of the objects. The main objective of this paper is providing precise recognition of objects and estimation of their location from an unknown scene. Whenever the object is recognized from extracted frames of the input video the background subtraction will be applied. Then the classification of the objects into their respective categories can be achieved using support vector machine classifier by supervised learning. In case of multiple objects of different classes in a single frame, a vector containing the classes of all the detected in that frame is produced as output. The results of this work are drawn in the MATLAB tool by considering the input video dataset taken from various sources and extracting the frames from the input video for the detection then the efficiency of the proposed techniques will be measured.

show abstract

Section: Related Workmentioning

confidence: 99%

Object Recognition using Lucas-Kanade Technique and Support Vector Machine Based Classification in Video Surveillance Systems

Pillai,

2019

IJEAT

View full text Add to dashboard Cite

show abstract

“…Detecting events and tracking situations is performed via various methods depending on the type of data (Deviatkov and Lychkov, 2017;Lychkov, Alfimtsev and Sakulin, 2018). To detect events in a stream of text documents, the authors proposed a method based on incremental clustering (Andreev, Berezkin and Kozlov, 2017).…”

Section: Detecting Events In Data Streamsmentioning

confidence: 99%

Forecasting of the Emergence and Development of Innovative Technologies

Proletarsky¹,

Andreev²,

Berezkin³

et al. 2019

Proceedings of the 16th International Conference on Applied Computing 2019

View full text Add to dashboard Cite

Under "Industry 4.0" conditions, the automatic prediction of emergence and evolution of innovative technologies is an important challenge. To solve this problem, a hybrid approach to big data streams analysis is proposed and developed. This approach allows the automated monitoring and forecasting of situations development through processing streams of heterogeneous data which is presented, in particular, by text documents, time series and database records. The proposed approach includes detecting events in data streams, forming situations, identifying possible scenarios for their further development and preparing suggestions for decision makers. To represent events that are reflected in a stream of structured data, a frame-based model is proposed. To form situational chains that reflect the development of innovative technologies, an incremental clustering method is used. Forecasting further evolution of the analyzed innovative technology consists in forming possible scenarios for the situation's development based on the case-based approach. An example of an event and a situation detected in a stream of structured data, as well as an example of a scenario generated for the detected situation are presented.

show abstract

“…The ill-posed definition of the visual tracking (i.e., model-free tracking, on-the-fly learning, single-camera, 2D information) is more challenging in complicated real-world scenarios which may include arbitrary classes of target appearance and their motion model (e.g., human, drone, animal, vehicle), different imaging characteristics (e.g., static/moving camera, smooth/abrupt movement, camera resolution), and changes in environmental conditions (e.g., illumination variation, background clutter, crowded scenes). Although traditional scenarios visual tracking methods utilize various frameworks -like discriminative correlation filters (DCF) [21]- [24], silhouette tracking [25], [26], Kernel tracking [27]- [29], point tracking [30], [31], and so forth -these methods cannot provide satisfactory results in unconstrained environments. The main reasons are the target representation by handcrafted features (such as the histogram of oriented gradients (HOG) [32] and Color-Names (CN)) [33] and inflexible target modeling.…”

Section: Introductionmentioning

confidence: 99%

Deep Learning for Visual Tracking: A Comprehensive Survey

Marvasti-Zadeh,

Cheng,

Ghanei-Yakhdan

et al. 2019

Preprint

View full text Add to dashboard Cite

Visual target tracking is one of the most sought-after yet challenging research topics in computer vision. Given the ill-posed nature of the problem and its popularity in a broad range of real-world scenarios, a number of large-scale benchmark datasets have been established, on which considerable methods have been developed and demonstrated with significant progress in recent yearspredominantly by recent deep learning (DL)-based methods. This survey aims to systematically investigate the current DL-based visual tracking methods, benchmark datasets, and evaluation metrics. It also extensively evaluates and analyzes the leading visual tracking methods. First, the fundamental characteristics, primary motivations, and contributions of DL-based methods are summarized from six key aspects of: network architecture, network exploitation, network training for visual tracking, network objective, network output, and the exploitation of correlation filter advantages. Second, popular visual tracking benchmarks and their respective properties are compared, and their evaluation metrics are summarized. Third, the state-of-the-art DL-based methods are comprehensively examined on a set of well-established benchmarks of OTB2013, OTB2015, VOT2018, and LaSOT. Finally, by conducting critical analyses of these stateof-the-art methods both quantitatively and qualitatively, their pros and cons under various common scenarios are investigated. It may serve as a gentle use guide for practitioners to weigh on when and under what conditions to choose which method(s). It also facilitates a discussion on ongoing issues and sheds light on promising research directions.

show abstract

Tracking of Moving Objects With Regeneration of Object Feature Points

Cited by 13 publications

References 12 publications

Object Recognition using Lucas-Kanade Technique and Support Vector Machine Based Classification in Video Surveillance Systems

Object Recognition using Lucas-Kanade Technique and Support Vector Machine Based Classification in Video Surveillance Systems

Forecasting of the Emergence and Development of Innovative Technologies

Deep Learning for Visual Tracking: A Comprehensive Survey

Contact Info

Product

Resources

About