Applications of Video Segmentation - Video Segmentation and Its Applications - page 149

Fig. 6.2 Finding
Vehicles [ 28 ]

tation, under segmentation (border holes), Flickering etc. The detection and false
alarms rates are estimated by counting how many times interesting and irrelevant
regions are detected.

6.4 Object-Based Scalable Video Coding

Surveillance centric methods reduce video bit rate without jeopardising informa-
tion relevant to surveillance application. As stated in [ 16 ], H.264/MPEG-4 AVC
provides a fully scalable extension, SVC, which achieves significant compression
gain and complexity reduction when scalability is sought, compared to the previous
video coding standards. Krutz et al. has described a methodology to separate a video
scene into shots that are coded either with an object-based codec or the common
H.264/AVC [ 11 ]. Their strategy is to use different video codecs for different kinds
of content in order to obtain higher coding gain. Contrasting conventional coders,
the system [ 22 ] addresses the requirements of surveillance application scenarios.
It aims at achieving bit-rate optimization, and adaptation of surveillance videos
for storing and transmission purposes. In the system, the encoder communicates
with a Video Content Analysis module that detects events of interests in videos
captured by CCTV. Bit-rate optimization and adaptation is achieved by exploiting
scalability properties of the target codec. Temporal segments containing events