Game Development Reference
In-Depth Information
2. High-efficiency and low-complexity video coding technologies should be adopted
in the system. To achieve higher compression rate, the special characteristics
of surveillance videos (e.g., the relatively stationary background in the scene)
should be utilized in the bitstream compression and coding. To make things better,
semantic coding in the object and event layers could be used to support the
structuring and summarization of surveillance video clips for both long-term
storage and retrieval.
The two requirements only can be solved at the edge, where the original videos
exist. By performing some basic analysis operators, the interesting objects or events
can be allocated with more bits and transmitted to analysis modules. Then, the load
on the infrastructure and people involved falls dramatically. A number of software
vendors supply intelligent video applications that solve specific needs. Together with
network cameras, video encoders and/or video management software systems, these
intelligent video applications, tailored to specific market requirements, have formed
complete solutions.
10.3.2 An Example Solution of Intelligent Video Coding
System—IEEE 1857
Nevertheless, it is not an easy job to empower an intelligent video coding system to
deal with all the problems. Fortunately, the recent IEEE 1857 standard which can
help the developers to construct such an intelligent video system is exactly a good
choice.
IEEE 1857, Standard for Advanced Audio and Video Coding, was originally
developed as early as 2006, and formally proposed to the IEEE Standards Association
in March 2012. The proposal was supported by Standards Activities Board of IEEE
Computer Society. The IEEE 1857 standard document was approved in March 2013,
and then released as IEEE 1857-2013 in June 2013. The most significant feature of
IEEE 1857-2013 is its special coding tools for surveillance video, which can achieve
at least twice coding efficiency on surveillance videos as AVC/H.264 HP, yet with a
slightly additional encoding complexity. Moreover, it is highly recognition-friendly
in that ROIs of attended objects can be encoded into the coded bitstream and thus
can be directly extracted for further analysis and recognition. These features make
the surveillance groups of IEEE 1857 one of enabling video coding standards and
tools for the next-generation smart video surveillance systems. In the rest of this
subsection, our discussion mainly focuses on IEEE 1857 surveillance groups.
Double the surveillance video coding efficiency
Different with other state-of-the-art standards such as HEVC/H.265 which
improves the coding efficiency at the cost of significantly increased coding com-
plexity, IEEE 1857 surveillance groups are mainly motivated by the fact that most
of surveillance videos are often captured by stationary cameras that always stand
toward the same scene for a long time. That is, there are usually similar background
