The Pet-Device++ Project will focus on those objectives:

• Video and Image Processing Parts
  
• Distributed Multimedia Databases based on Web Service.
  
• MPEG-7 Ontology for semantic representation of media content.

Image & Video Processing Part

Image Retrieval for MPEG-7 visual descriptor

• Efficient image retrieval using color descriptors.
• Multi-feature image retrieval: combine color and texture descriptors to retrieve.
  
• Region-based image retrieval.

Video Segmentation and Annotation

• Video Segmentation utilizing Visual and Audio Features
  
• Automatic Annotation by exploring the semantic content in the video data.
  
• Video Retrieval based on the affiliated annotated video information.

Video Motion Analysis

• Shot Boundary Detection.
  
• Camera Motion Estimation.
  
• Extraction Motion Feature.
  
• Build Motion Class Model.
  
• Do Motion Classification.
  

Distributed Multimedia Databases

Creating Multimedia Databases from XML file.

• Creating MPEG-7 Description source database.
  
• Mapping all the descriptors provided by MPEG-7 standard into Relational Database.
  
• Decomposing existing MPEG-7 Description files into the source database and composing the
• MPEG-7 description documents from the source database

Association Mining in XML

• focus on the change association among substructures when semistructured objects evolve.
  
• Two kinds of structure are considered: inter-structure and intra-structure
  
• Mining associated substructures in a collection of different versions of XML or other semistructured data.

Extension for UDDI

• Provides QoS-Aware and Federated Support for UDDI.
  
• The architecture is aware of the provider services’ QoS.
  
• The measurement are shared.
  
• Customization is available in the discovery procedure.
  
• The extended inquiry interface conforms to the UDDI V2 specification.
  
• The system supports the discovery across federated domains.
  

MPEG-7 Ontologies for Semantic representation of media content

Image Retrieval for MPEG-7 visual descriptor

• Focus on browsing and retrieval multimedia contents efficiently and accurately.
  
• Bridge high-level human conceptual items to machine-processable data though building theontologies for high-level multimedia content descriptions.
  
• Endow retrieval system with certain intelligence to deal with inference and fuzzy retrieval.
  

Faculty Members
Dr. Lee Bu Sung (Assoc. Prof)
Dr. Chia Liang Tien, Clement (Assoc. Prof)
Dr. Deepu Rajan (Asst. Prof)
Dr. Sourav S Bhowmick (Assoc. Prof)
Dr. Chng Eng Siong (Asst. Prof)

Research Scholars
Ms. Wang Surong (Ph.D)
Ms. Chen Ling (Ph.D)
Mr. Zhou Chen (Ph.D)
Mr. Liu Song (Ph.D)
Mr. Chu Yang (Ph.D)

Contact Information
Assoc. Prof. Chia Liang Tien, Clement
Director, Centre for Multimedia and Network Technology
School of Computer Engineering
Nanyang Technological University
N4-#B2C-06, Nanyang Avenue
Singapore 639798
Tel:(65)6790-5832