Computer Vision and Image Understanding: Large Scale Multimedia Semantic Indexing
Computer Vision and Image Understanding Call for Papers
Special Issue on Large Scale Multimedia Semantic Indexing
Scope
The ever-increasing huge volume of multimedia data in Web-sharing
sites and personal archives has provided us both challenges and unique
opportunities on massive multimedia management. Due to the well-known
semantic gap between human-understandable high-level semantics and
machine-generated low-level features, recent years have witnessed
plenty of research effort on large-scale multimedia content
understanding and indexing. This special issue aims to collect recent
state-of-the-art achievement on multimedia semantic indexing,
especially the work devoted to several new challenges in this
field. For example, it is recently discovered that facilitated with
the contextual information in social multimedia, concept detection in
videos can be better accomplished. Other problems such as cross-domain
concept detection and multi-modality semantic learning are also in the
interest of this special issue. Moreover, due to the explosive
increase of both the size of the multimedia database and feature
dimension, it is highly desired that an algorithm for semantic
indexing or multimedia retrieval can be applied in a large-scale
setting. An ideal algorithm should be a good balance between
effectiveness and computation efficiency. Another focus of this
special issue will be on recent advances on scalable
algorithms. Particularly, locality-sensitive hashing (LSH) method
which originates from theoretic computer science has attracted
extensive research interest in the past years. Its success has been
demonstrated in various applications, such as near-duplicate image
detection. However, it is still an open problem how the hashing
algorithm can be most effective, given the various complications in
multimedia data, including diverse multimedia semantics, specific
intrinsic data structure (e.g. graph or low-dimensional manifold) and
multi-modality features. The special issue target at collecting latest
research breakthroughs from both theoretic study and the related
applications. Novel semantic indexing algorithms that are capable of
handling large-scale data are highly appreciated. Inspiring work that
discusses promising future directions is also welcome. This special
issue targets the researchers and practitioners from both the industry
and academia. Topics of interest include but not limited to:
Ontology design and semantic concept detection
o Lexicon of semantic concepts
o Concept detection and semantic attribute extraction
o Novel feature representation and semantic indexing for image, audio
and video data
o Cross-domain concept learning
o Semantics-oriented image and video annotation
o Fusion methods for multi-modality features
o Novel machine learning techniques for semantic features
o Concept detection in social multimedia
Large scale semantic indexing algorithm
o Locality-sensitive hashing for multimedia semantic representation
o Hashing for complicated data structures (e.g., graphs, manifolds,
multiple-instance data)
o Hashing in kernel space
o Hashing for multi-modality representation
o Benchmarks and evaluations for multimedia hashing
Related applications
o Large-scale image or video retrieval
o Multimedia event detection
o Image annotation / tagging / recognition
o Query-adaptive methods for multimedia retrieval and event detection
o Large scale cross-media retrieval
Important Dates:
Paper submission due: Aug. 15, 2013
First notification: Nov. 15, 2013
Revision: Dec. 30, 2013
Final decision: March. 15, 2014
Publication date: Autumn 2014 (Tentative)
Guest Editors:
Dr. Yadong Mu, Columbia University, USA (muyadong@gmail.com)
Dr. Yi Yang, Carnegie Mellon University, USA (yiyang@cs.cmu.edu)
Dr. Liangliang Cao, IBM T.J. Watson Research Lab, USA
(liangliang.cao@us.ibm.com)
Prof. Shuicheng Yan, National University of Singapore
(eleyans@nus.edu.sg)
Prof. Qi Tian, University of Texas at San Antonio, USA
(qitian@cs.utsa.edu)