A multi-Kalman filtering approach for video tracking of human-delineated objects in cluttered environments

Gao, Jean; Kosaka, Akio; Kak, Avinash C.

doi:10.1016/j.cviu.2004.10.005

Cited by 22 publications

(6 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The proposed ARMS 2 NOT approach is realized to find the exact location of the desirable number of objects in all frames of a video sequence inspired by the basic concept of mean shift tracking approach, as one of well-known algorithms, in the area of object-tracking methods (Gao et al, 2005). All of the chosen objects are here needed to be tracked, in a synchronized manner, without a lengthy processing time.…”

Section: The Primary Materialsmentioning

confidence: 99%

“…In such a case, as long as occlusion does not occur, the SCGMM learns for each object, where the learned SCGMMs are suitable for the segmentation of the current occlusion, since a displacing procedure is utilized for adapting the SCGMMs to the spatial variations. Gao et al (2005) have also suggested a multi-Kalman filtering approach for video tracking of human-delineated objects in cluttered environments. In the Gao work, a approach that uses a motion-estimation based framework for video tracking of objects in cluttered environments has been proposed.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multiple stationary–non-stationary object-tracking approach in real-time applications through an extendable RGB modelling framework

Kazemi

Mazinan

2013

Transactions of the Institute of Measurement and Control

View full text Add to dashboard Cite

The main contribution of the present research arises from developing standard object tracking approaches by considering a number of available research works in the area of multiple-object tracking approaches in real-time applications. An automatic robust algorithm is proposed in this investigation, which is able to track multiple synchronized stationary and/or non-stationary objects in high performance quality. The proposed approach is realized based upon an efficient extendable RGB (red/green/blue) modelling framework. This means that the investigated results are easily applicable to tracking the desirable number of objects to be chosen, i.e. the approach is flexible enough to reorganize to track the stationary/non-stationary objects in all frames of the video sequences. This proposed idea unique in that we plan to track all the chosen objects, in a synchronized manner, without a lengthy processing time. In fact, the proposed approach is realized in association with a desirable number of sub-systems, each of them needing to be enabled in parallel. The investigated results indicate that the present approach is now applicable in real-time applications by running a robust estimator to predict the new position of the tracked objects, since the objects are in the presence of long-term environmental problems – situations that are usually too difficult to deal with through standard object tracking algorithms.

show abstract

Section: The Primary Materialsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Multiple stationary–non-stationary object-tracking approach in real-time applications through an extendable RGB modelling framework

Kazemi

Mazinan

2013

Transactions of the Institute of Measurement and Control

View full text Add to dashboard Cite

show abstract

“…The study by Rossi, Abderrahim and Diaz [7] shows that some of the most popular approaches are based on the use of Kalman filters. The Kalman filter and its extensions have been successful for moving vehicle tracking, but Gao, Kosaka, and Avinash [8] find that these methods do not cope with highly nonlinear models or non-Gaussian noise. Besides, Choi, et al [9] use template matching based on feature points for vehicle tracking.…”

Section: Introductionmentioning

confidence: 99%

“…are approximate and detailed coefficients of ( ) f t in m V respectively. According to Mallat Arithmetic [7,8], we have:…”

Section: Introductionmentioning

confidence: 99%

Vision-Based Vehicle Detection in Real Traffic Environment Using Fast Wavelet Transform and Kalman Filter

Zhuo

Lin

2014

AMR

View full text Add to dashboard Cite

A new vision-based approach for robust vehicle detection is addressed in this study. First, fast wavelet transform (FWT) is proposed to extract image texture, while grey level co-occurrence matrix (GLCM) is employed to measure and analyze the extracted texture. Then, vehicles can be extracted because the vehicle sections and the shadow sections have different textures in the foreground image. Moreover, we put forward the state and observation matrixes of Kalman filter which can be used to track vehicles under complicated traffic scenes. Experimental results in real traffic scenes show that the proposed methods are effective and efficient.

show abstract

“…Filtros estocásticos são conhecidos e amplamente usados em problemas de rastreamento [7,8,9,2]. Eles serão apresentados posteriormente, quando serão modificados e usados na tarefa de fusão de dados.…”

Section: Rastreamento De Contornosunclassified

Rastreamento do contorno do ventrículo esquerdo em imagens de ressonância magnética usando filtragem estocástica

Richa¹

View full text Add to dashboard Cite

Agradeço inicialmente aos meus pais pelo apoio constante durante todo meu curso de graduação longe de casa e aos meus tios que me acolheram como um filho, me provendo todo suporte necessário. Gostaria de agradecer a dedicação, paciência do incansável professor Geovany Araújo Borges, pela base fundamental à minha formação como engenheiro. Agradeço também à todos os colegas do GPDS, que sempre colaboraram com meus trabalhos e em especial ao professor Francisco Assis que me permitiu a entrada no grupo. Finalmente, agradeço o carinho de todos amigos por todos momentos de alegria durante estes últimos anos. Rogério de Almeida Richa RESUMO A segmentação e rastreamento do contorno do ventrículo esquerdo em imagens de ressonância magnética é um trabalho desafiador. O contorno do ventrículo esquerdo (VE) oferece dados importantes sobre o comportamento do coração, sendo útil para descoberta de disfunções, para o cálculo de parâmetros como fluxo sangüíneo que circula no coração etc. Para tornar o rastreamento do contorno do VE robusto à ruído, baixo contraste e outras complicações, filtros estocásticos são propostos para modelar as incertezas na localização do contorno real nas imagens. O algoritmo proposto neste trabalho realiza, em tempo real, o rastreamento em uma seqüência de imagens de IRM compreendendo diversos ciclos cardíacos. Seguindo uma abordagem baseada em contornos ativos, pouco conhecimento sobre a deformação das paredes do ventrículo é requerido para a realização do rastreamento. Finalmente, o método proposto foi avaliado usando uma seqüência de imagens cedida por pesquisadores da University of Southern of California (USC).

show abstract

A multi-Kalman filtering approach for video tracking of human-delineated objects in cluttered environments

Cited by 22 publications

References 39 publications

Multiple stationary–non-stationary object-tracking approach in real-time applications through an extendable RGB modelling framework

Multiple stationary–non-stationary object-tracking approach in real-time applications through an extendable RGB modelling framework

Vision-Based Vehicle Detection in Real Traffic Environment Using Fast Wavelet Transform and Kalman Filter

Rastreamento do contorno do ventrículo esquerdo em imagens de ressonância magnética usando filtragem estocástica

Contact Info

Product

Resources

About