An efficient parallel algorithm for motion estimation in very low bit-rate video coding systems

Abstract

SUMMARYMotion estimation is widely used in video coding schemes in order to reduce the inherent temporal redundancy among the frames of a video stream. In particular, low and very low bit rate video coding schemes need sophisticated motion models which usually require a large number of arithmetic operations. In this paper we present a parallel algorithm for the most practical of these models. Specifically we implement the affine motion model on a hypercube-based multiprocessor. This model covers the most usual kinds of motion and requires only a modest number of arithmetic operations. Also, the hypercube network can efficiently handle the non-regular data flow resulting from the parallel implementation of this model. In addition, we assume that our multiprocessor is fine grained, in contrast to most programmable architectures used in video coding, where processors usually have large local memory. Apart from its practicality, the constraint of limited local memory makes the algorithm design more challenging and thus more theoretically interesting. Finally, with regard to other proposals in the literature, our scheme is more general: whereas our scheme covers all kinds of motion supported by the affine motion model, the rest of the proposals deal only with a subset of these kinds.

A parallel algorithm for motion estimation in video coding using the bilinear transformation

“…The rich interconnection structure of this network favours the design of “elegant” parallel algorithms for a number of problems (Grama et al 2002 ) which can be used in other parallel models (Sundar et al 2013 ), as well. Following the basic methodology of (Konstantopoulos et al 2000 ), we present a hypercube algorithm with low communication and computation cost. We formally prove those good features and we also analytically determine the memory required per processor for running the algorithm.…”

“…In a previous work (Konstantopoulos et al 2000 ), we have designed a parallel algorithm on the parallel model of the hypercube network for motion estimation in video using the affine transformation. We have demonstrated how to perform this estimation with low time complexity as well as with low local memory requirements per processor.…”

“…We have demonstrated how to perform this estimation with low time complexity as well as with low local memory requirements per processor. In this paper, we follow the general methodology in (Konstantopoulos et al 2000 ), and present a parallel motion estimation algorithm based on the bilinear transformation again on the hypercube. Note that the bilinear transformation is more complex than the affine one since the latter is a special case of the former.…”

Παράλληλες Υλοποήσεις Θεμελιωδών Εργαλείων Της Τεχνολογίας Των Πολυμέσων

“…Σε σχέση με άλλες προτάσεις που έχουν εμφανιστεί στη βιβλιογραφία, ο σχεδιασμός μας είναι πιο γενικός: ενώ ο δικός μας σχεδιασμός καλύπτει όλα τα είδη κίνησης που περιγράφονται από τον affine μετασχηματισμό κίνησης, οι υπόλοιπες προτά σεις για το ίδιο θέμα χειρίζονται μόνο ένα υποσύνολο από αυτά τα είδη κίνησης. Τα αποτελέσματα αυτής της μελέτης εμφανίστηκαν στα [50,51]. Στο δεύτερο τμήμα αυτού του κεφαλαίου παρουσιάζουμε συνοπτικά ένα παράλληλο αλγόριθμο για τη περίπτωση που ο υπολογισμός κίνησης των αντικειμένων σε μία ακολουθία εικόνων γίνεται με βάση το bilinear μετασχηματισμό κίνησης.…”