Visual Tracking combining Region Covariance with Gaussian Process Prediction

Título: Visual Tracking combining Region Covariance with Gaussian Process Prediction

Autores: Marques, Humberto;Chaves, Levy;Bessa Maia, José Everardo

Resumo:
This work implements a video tracking method which combines region covariance detection with motion estimation by Gaussian Process Regression. The computational and accuracy performance of the proposed method is compared to two other state-of-the-art algorithms available in OpenCV using benchmark videos from the VOT public challenges. The results show that the method is competitive and can be used in the construction of practical applications in several contexts of video surveillance.

Palavras-chave:
Visual Tracking;Region Covariance Model;Gaussian Process Prediction.

Páginas: 10

Código DOI: 10.21528/CBIC2017-103

Artigo em pdf: cbic-paper-103.pdf

Arquivo BibTeX: cbic-paper-103.bib