Game Development Reference
In-Depth Information
Fig. 12.8
Illustration of the
variable size of coding unit
in HEVC
can adaptively change with the object boundary shape. Generally more details with
object boundaries lead to finer block partitions, as illustrated in Fig.
12.8
.
The visual object-based coding has shown its efficiency in the applications of
the surveillance video compression and drawn many research attentions in recent
years. In the surveillance video compression, the foreground objects are extracted by
a low-complexity background modeling method (Zhang et al.
2014
) and compressed
with more accurate prediction achieving up to 50% bitrate saving comparing with
the newest HEVC standard. In Huang et al. (
2014
), a scheme for representing visual
objects in HEVC coding loop is proposed, where three methods are designed. The
first two methods, dubbed as flagging compression blocks (FCB) and additional
object tree (AOT), respectively, introduce some additional bits to the coding blocks
for reconstructing the object boundary. Both of the two techniques would not change
the coding procedure. The third method referred to as confining by shape (CBS) is
quite different because the split decision of coding unit is guided by the object shape,
and it will inversely influence the coding loop. As such the representation of visual
objects can be implemented in the HEVC loop by reusing the variable-size blocks
without any extra tools.
Overall, at the background of big data, the visual object-based coding has been
revisited again and drawn more and more attentions. Many efforts has been done
for these specific applications such as surveillance video compression. For general
video coding, some initial works have been presented for efficient representation of
visual objects in the traditional coding loop. In the future, the conventional video
compression and some novel ideas including the object based representation can be
combined together to achieve more efficient compression.
12.4 Future Challenges
12.4.1 Quality of Service-Based Video Compression
Currently, there are a lot of discussions on the quality of service (QoS) of the multi-
media communication network. The ultimate receiver of the multimedia, including
the audio, images and videos, are human beings. Therefore, the QoS issues shall
