Constrained optimization techniques provide an effective solution to the rate-distortion optimized mode selection in hybrid motion compensation-DCT based video encoders. However, the minimization of the Lagrangian cost function often presents a computational burden in real-time applications. In this paper, we propose a new method for rate distortion optimized coding mode selection. Instead of using the Lagrangian method, we employ a computationally efficient, yet highly effective rate-distortion model to minimize the overall distortion of a macroblock. Extensive experimental results show that the proposed method provides PSNR improvements of 2.79, 1.64 and 1.54 dB over TM5-based MPEG-2 encoder, at the bit-rates of 2, 4 and 6 Mbps, respectively, with negligible increase in computational complexity.
Published in:
Multimedia and Expo, 2003. ICME '03. Proceedings. 2003 International Conference on
(Volume:3
)
Date of Conference: 6-9 July 2003