Title: Towards a distributed and parallel schema for active appearance model implementation
Authors: Oum El Kheir Naoui; Saïd Mahmoudi; Ghalem Belalem
Addresses: Department of Computer Science, Faculty of Exact Sciences and Applied, University of Oran, Algeria ' Computer Science Department, Faculty of Engineering, University of Mons, Mons, Belgium ' Department of Computer Science, Faculty of Exact Sciences and Applied, University of Oran, Algeria
Abstract: In this paper, we propose a distributed and parallel schema to implement the standard version of active appearance model segmentation approach proposed by Cootes. In the scheme distributed, we project to integrate the parallelism to implement the model operations. The parallel schema proposed is based on the data parallelism. The majority of operations of the model lend themselves well to the parallelism. We quote the alignment procedure and the calculation of intensity profiles. These two operations belong to the stage of modelling shape. The proposed schema is intended to exploit the benefits built in the area of parallel and distributed image processing. We theoretically demonstrate the feasibility of the parallel approach for the first model operations, in order to generalise this approach to other operations. This implementation is still being and is aiming to exploit the benefits of the recent software tools such as Java-RMI, and parallel programming.
Keywords: active appearance model; segmentation; parallel image processing; distributed image processing; data parallelism; modelling.
DOI: 10.1504/IJCVR.2016.073756
International Journal of Computational Vision and Robotics, 2016 Vol.6 No.1/2, pp.19 - 40
Received: 31 Dec 2013
Accepted: 27 Sep 2014
Published online: 18 Dec 2015 *