PR

Pattern Recognition

Video Analytics with Deep Learning

We are living in a world where we are surrounded by so many intelligent video-capturing
devices. These devices capture data about how we live and what we do. For example,
anks to surveillance and action cameras, as well as smart phones and even old-fashioned
amcorders, we are able to record videos at an unprecedented scale and pace. There is
exceedingly rich information and knowledge embedded in all those videos. With the recent
advances in computer vision, we now have the ability to mine such massive visual data to
obtain valuable insight about what is happening in the world. Due to the remarkable
successes of deep learning techniques, we are now able to boost video analysis performance
significantly and initiate new research directions to analyze video content. For example,
convolutional neural networks have demonstrated superiority on modeling high-level visual
concepts, while recurrent neural networks have shown promise in modeling temporal dynamics
in videos. Deep video analytics, or video analytics with deep learning, is becoming an
emerging research area in the field of pattern recognition.

The goal of this special issue is to call for a coordinated effort to understand the
opportunities and challenges emerging in video analysis with deep learning techniques,
identify key tasks and evaluate the state of the art, showcase innovative methodologies
and ideas, introduce large scale real systems or applications, as well as propose new
real-world datasets and discuss future directions. The video data of interest cover a
wide spectrum, ranging from first-person wearable videos, web videos (aka user-generated
content), commercial video programs, to surveillance videos. Video analytics plays an
important role in public security, entertainment, healthcare, and so on. We solicit
manuscripts in all fields of video analytics that explore the synergy of video understanding
and deep learning techniques.

We believe the special issue will offer a timely collection of research updates to benefit
the researchers and practitioners working in the broad computer vision and pattern recognition
communities. To this end, we solicit original research and survey papers addressing the topics
listed below (but not limited to):

Topics:

o First-person/wearable video analysis using deep learning techniques, including object
detection and recognition, highlight detection, action recognition, event detection,
segmentation and tracking, classification, summarization and storytelling, editing, data
collection and benchmarking, and so on.
o Video and language – describing video with natural language using deep learning techniques.
o Web video understanding using deep learning techniques, including classification, annotation,
event detection and recognition, authoring and editing, and summarization.
o Home/public video surveillance using deep learning, including motion detection and classification,
scene understanding, event detection and recognition, people analysis, object tracking and
segmentation, human computer/robot interaction, behavior recognition, crowd analysis, fusion of
vision with other sensing modalities, and so on.
o Data collections, benchmarking, and performance evaluation.