Abstract
We propose a transcoding method of H.264 coded bitsteam to control the picture quality dependently on the interest region. In the proposed method, first we find the model of quantization step-size and bitrate. Then a classification method according to the subjectively interest region within a video sequence is suggested. Also we propose a method that assigns a specific quantization step-size differentially according to the interest region within a video. In general, the subjective picture quality can be increased by applying the quantization step-size as a small value relatively for the interest region compared with the other regions. From the simulation, we show that the proposed method can make better subjective picture quality relatively in parts of interest region.
Highlights
INTRODUCTIONTHE conversion of the bitrates with keeping the same coding standard or syntax is widely used for various applications in the inter-network communications such as video streaming services
THE conversion of the bitrates with keeping the same coding standard or syntax is widely used for various applications in the inter-network communications such as video streaming services.When a final user cannot accept the higher bitrate of the originally compressed or/and stored bitstream, the bitrate reduction may be needed
For an MPEG-1,2 encoding system, we have proposed a formula based on a linear relation between bitrate (R) and log function of a quantization step-size (Qstep) in a small range of bitrate [10]
Summary
THE conversion of the bitrates with keeping the same coding standard or syntax is widely used for various applications in the inter-network communications such as video streaming services. For reducing the video bitrate, the transformed coefficients are re-quantized with a larger quantization step-size. The coefficients generated by the re-quantization are converted into a new bitstream with a lower bitrate after the variable length code encoding process. This bitstream transcoding operation does not include the re-ordering and re-estimation for the motion compensation. All of methods concern with proper allocation of bitrate or quantization step-size in order to satisfy good picture quality or high PSNR (Peak Signal to Noise Ratio).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
