Exploiting depth information for fast multi-view video coding

Abstract

This research work is partially funded by the Strategic Educational Pathways Scholarship Scheme (STEPS-Malta). This scholarship is partly financed by the European Union –\ud European Social Fund (ESF 1.25).Multi-view video coding exploits inter-view redundancies to compress the video streams and their associated depth information. These techniques utilize disparity estimation techniques to obtain disparity vectors (DVs) across different views. However, these methods contribute to the majority of the computational power needed for multi-view video encoding. This paper proposes a solution for fast disparity estimation based on multi-view geometry and depth information. A DV predictor is first calculated followed by an iterative or a fast search estimation process which finds the optimal DV in the search area dictated by the predictor. Simulation results demonstrate that this predictor is reliable enough to determine the area of the optimal DVs to allow a smaller search range. Furthermore, results show that the proposed approach achieves a speedup of 2.5 while still preserving the original rate-distortion performance.peer-reviewe

Fast disparity estimation for Multi-view plus depth video coding

“…The performance of this technique is investigated for both color and depth MVC of two different MVD sequences. A speedup gain of 21.8 is achieved over the original DE while a speed-up of 3 is achieved over our previous work in [9], for MVC of both color and depth MVVs.…”

“…Our previous work [9] showed that, for DE, the multiview geometry together with the average depth data can be used to identify the location of the current sub-MB's top-left corner from the Target frame to the 3D space. Its inverse is then used to re-localize it back in the view Reference frames.…”

“…However, in MVD coding, the full geometrical information provided by the available depth data, can be used to identify a more accurate and a smaller search area for DE in both color and depth MVC. Our previous work in [9] demonstrated that multi-view geometry together with depth data can be used to identify accurate initial disparity search points. This allows the DE's search area to be limited to a small fixed square area.…”

Exploiting depth information for fast motion and disparity estimation in Multi-view Video Coding

“…When analyzing the results obtained by the MVC with both the proposed geometric estimations, with respect to using only the proposed DE in [11], it can be noted that the novel solution provides a further speed-up gain of 1.7 times for the FSE method and of 1.4 times for the FASE method. This speed-up gain was attained without any significant increase in the CODEC's complexity, thus, the latter is preferred.…”

“…The search area can be reduced since this predictor gives a very good estimate of the optimal DVs. A similar method was already reported in our previous work [11] and can still be used to reduce the DE's computations.…”