|
Computer Vision is the science and technology of machines that see. The dominant
scientific conferences in computer vision, such as ICCV, CVPR and ECCV,
concentrate on theories and models for obtaining information from images and
image sequences. The intensely competitive nature of these conferences leaves
little room to address the systems and engineering science issues necessary for
transforming vision theory into practical vision systems. The International Conference
on Vision Systems, ICVS, was created to fill this gap.
The first ICVS was organized in December 1999 at Las Palmas in the Canary
Islands to provide a forum for research on systems architecture, benchmarks and
performance evaluation. Since that time, the field of computer vision has made
impressive progress, with the emergence of reliable techniques for interest point
detection and matching, image indexing, category learning, object detection,
recognition and classification. Meanwhile, computing power has become much
less of a barrier to building vision systems. Desktop computers have evolved from
machines with 100-MHz clocks and a few megabytes of memory to multi-core
architectures with multiple GHz clock processors and gigabytes of memories.
This progress has been reflected in the emergence of techniques documented in
ICVS conferences in Toronto in 2001, Vancouver in 2002, Graz in 2003, New
York in 2005, Santorini in 2007, and Li`ege in 2009. We continued this tradition
with the 8th International Conference on Vision System in Sophia Antipolis.
The conference Program Committee received 58 submitted papers. Each paper
was assigned to three reviewers from among the 31 members of the review
committee, leading to the selection of 22 papers for oral presentation at the conference.
These were organized into seven sessions, showcasing recent progress in
the areas performance evaluation, activity recognition, control of perception, and
knowledge-directed vision. The program was completed by presentations from
three invited speakers, exploring areas of particularly high potential for impact
on the engineering science of vision systems.
The emergence of mobile computing has led to a revolution in computer vision
systems. The ubiquitous nature of cameras on mobile telephones and tablets
has enabled new applications that combine vision and mobility with ubiquitous
access to information over the Internet. However, the limited computing and
electrical power of mobile platforms has limited these systems. This is set to
change with the emergence of low-power GPUs specifically designed to support
computer vision and graphics on mobile devices. The invited talk by Joe Stam of
NVIDIA described the emerging use of GPUs as a hardware platform for vision
systems on personal computers and described the new generation of devices for
mobile platforms such as cameras and tablets. |