RATE CONTROL TECHNIQUE BASED ON 3D QUALITY OPTIMIZATION FOR 3D VIDEO
ENCODING
Wen-Nung Lie
1,2
and Yu-Peng Liao
1
1
Department of Electrical Engineering, National Chung Cheng University, Taiwan, ROC.
2
Advanced Institute of Manufacturing with High-tech Innovations (AIM-HI), National Chung Cheng
University, Taiwan, ROC.
{ieewnl@ccu.edu.tw}
Abstract —
This paper presents a new 3D video (in color
plus depth format) encoding system, featured of 3D quality
optimization and joint rate control between color and depth
components. An SVR-based prediction model is pre-built
for estimating the best bit rate allocation between color and
depth for each frame by analyzing edge features in the
color and the depth component images. We also modify
rate control scheme in H.264/SVC JSVM reference
software to be suitable for joint rate control of the color and
depth sequences. Accordingly, bit rates can be dynamically
and accurately allocated between color and depth
components to achieve optimized 3D video quality. From
the experiment results, it is shown that our proposed
method is capable of achieving more accuracy in rate
control and better 3D quality for visual perception.
Keyword ─ 3D video, 3D quality, rate control, support
vector regression (SVR) model
I. I
NTRODUCTION
With the fast development of display technology, 3D
video is capable of providing observers a more immersive
and better visual experience. Compared with the traditional
2D video, 3D video contains much more data that are
required to provide depth perception effect for observers.
Recently, both the stereo and multi-view high profiles have
been standardized in JMVC (Joint Multi-View Video
Coding), whereas another format, multi-view video plus
depth (MVD), which includes information about per-pixel
scene geometry for each view, has also been discussed
intensively for 3D video. This video plus depth (V+D)
format is advantageous of multiple-view (up to tens)
rendering for auto-stereoscopic display while it is encoded
and transmitted to the receiver at a cost of moderate
increase in bit-rate for the depth information.
Recently, a number of researches were focused on the
encoding of V+D 3D video. The color and depth parts are
either separately encoded (e.g., in terms of two H.264/AVC
video streams) or jointly encoded [13] (e.g., based on the
H.264/SVC (Scalable Video Coding) platform). However,
few of them address the issue of rate allocation between
color and depth. Color/depth rate allocation differs from
traditional rate control in video coding; it seeks for a
partition of bit rates for color and depth parts on a sequence
or frame (in this paper, it is “frame”) basis so that the
reconstructed 3D video quality can be optimized.
Intuitively, 3D video quality might be linked to the
reconstructed color and depth images. However, depth
images are only an intermediate media that are used for
synthesizing novel views (based on the Depth-Image-Based
Rendering, DIBR, technique) after being decoded at
receiver. Humans actually perceive 3D depth effect offered
by the synthesized novel views, but not the depth images.
How to control the distortion of depth during encoding
process such that the accordingly synthesized novel views
result in a good 3D quality constitutes an open issue in
recent researches.
It is naturally difficult to predict the quality of the
synthesized novel views when encoding a depth image
without performing the realistic DIBR process. However, a
number of DIBR processes will be required when adjusting
encoding parameters (e.g., quantization parameter, QP) for
the depth images to meet a target bit rate. Considering the
depth target bit rate as an unknown, the rate control
problem will certainly become much more complicated. It
is our goal in this paper to further resolve the optimization
problem of dynamic bit rate allocation between color and
depth parts when they are jointly encoded at a given total
bit rate.
A number of traditional researches were focused to
establish metrics for 2D/3D image quality assessment. The
metrics often rely on a set of per-pixel image features and
mapping of them to a score (e.g., a floating number
between 0 and 1) which presents a high positive correlation
with the human perception. However, few of them, except
the conventional PSNR metric, are linked to a video
encoder, i.e., adjusting parameters to optimize the resulting
encoded image quality.
Video rate control is meant to control the resulting bit
rate during encoding process to meet a target value, while
maintaining a better reconstruction quality simultaneously.
Traditional rate allocation/control techniques are based on
the PSNR or MSE metric to evaluate the resulting 2D
image quality. However, for the encoding of 3D
978-1-4799-5751-4/14/$31.00 ©2014 IEEE ICIP 2014140
资源评论
码流怪侠
- 粉丝: 3w+
最新资源
- FTLF1429P3BNV-高速SFP+光纤收发器-中文数据手册.pdf
- DS1100-5级经济型定时器-中文数据手册.pdf
- HDSP-521x系列-数码管-中文数据手册.pdf
- MM74HCT164-8位串入并出移位寄存器-中文数据手册.pdf
- HMC613LC4B-连续检测对数视频放大器-0.1GHz~20GHz-中文数据手册.pdf
- PESD5V0X1BL-超低电容双向ESD保护二极管-中文数据手册.pdf
- PI3USB102-USB2.0高速（480Mbps）信号开关-中文数据手册.pdf
- NSI45020AT1G-恒流调节LED驱动器-中文数据手册.pdf
- ANT1608LL14R2400A系列-表贴2.4GHz~2.5GH天线-中文数据手册.pdf
- TSOP393系列-红外遥控接收模块-中文数据手册.pdf
- SQ2361AEES-耐高温PMOS-中文数据手册.pdf
- BB208-02-低压可变电容二极管-中文数据手册.pdf
- 主题巴巴主题源码 空巴主题 合辑打包下载+主题巴巴SEO插件 WordPress主题模版网站源码.zip
- Java开发集合内存管理技巧与优化：合理选择与使用集合提升程序性能和避免内存泄漏Java开发中集合
- 中国省级、城市AI算力数据（2000-2024年）
- 【Java内存管理】集合框架优化实践：现代特性与大数据处理效率提升策略
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
