Segmentação de texturas dinâmicas: um novo método baseado em caminhadas determinísticas GONÇALVES, W. N. ; MACHADO, B. B. ; BRUNO, O. M. VII Workshop de Visão Computacional, 2011. v. 1, p. 73-78.
Recently, dynamic texture has attracted steadily growing interest due to the number of multimedia databases. Dynamic texture emerged as a new field of research that extends the concept of self-similarity of texture to the spatiotemporal domain. In this paper, we propose a new approach for dynamic texture segmentation based on deterministic walks and k-means algorithm. Although k-means presents interesting results, it only guarantees its convergence to a local minimum, which can affect the segmentation. To overcome this issue, this paper also presents a comparative study of six initialization methods of k-means applied to the dynamic texture segmentation problem. Experimental results on 300 databases demonstrate the effectiveness of the proposed method compared to the state-of-the-art methods.