3-D human shape analysis and retrieval for video surveillance

DEL PhD Studentship 2013/14

3-D human shape analysis and retrieval for video surveillance

Principal Supervisor: Drs. Paul Miller and Huiyu Zhou

Project Description:

3-D human body model recognition recently gains boosting attention due to the availability of commodity devices such as Microsoft’s Kinect. Large collections of 3-D human model are now easily available in the public, e.g. Internet. Therefore, analysing and retrieving these 3-D data collections have quickly become a topic of general interests in the areas of human computer interface and homeland security.

To search for a 3-D human model that looks similar to a query model, people normally apply the similarity of geometric distributions to describing the degree of similarity. This geometry based strategy has two groups: shape based and topology based matching. In spite of promising achievements, these two approaches have exhibited poor performance in the case of pose changes and/or background clutters.

The aim of this work is to develop a novel 3-D retrieval approach, based on sparse coding, which demonstrates a strong robust capability against human pose changes, illumination variations and clutters in the scene. The proposed retrieval system consists of fast feature extraction and on-line retrieval process. To improve the robustness of the system against rotation/scaling, we use the combination of multiple features, e.g. shape contexts, integral volume descriptors, heat Kernel signatures, and mesh histogram of gradients.

Objectives:

To investigate an appropriate combination of multiple descriptors in representing 3-D human body models.
To develop a new classification approach for 3-D human body models, based on sparse coding.
To optimise the developed retrieval system for efficiency and effectiveness.
To evaluate the entire retrieval system over publicly accessible databases.

Academic Requirements:

A minimum 2.1 honours degree or equivalent in Electrical and Electronic Engineering or Computer Science or other relevant subject is required.

General Information

This 3 year PhD studentship, funded by the Department for Employment and Learning (DEL), commences on 1 October 2013, covers approved tuition fees and a maintenance grant (unknown for 2013/14) is approximately £13,000 - £14,000.

Applicants should apply electronically through the Queen's online application portal at: https://dap.qub.ac.uk/portal/

Further information available at: http://www.qub.ac.uk/schools/eeecs/PhD/PostgraduateResearchScholarships/

Contact details:

Supervisor Name:	Dr. Huiyu Zhou
Address:	The Institute of Electronics, Communications and Information Technology (ECIT) Queen's University Belfast NI Science Park Queen's Road, Queen's island Belfast, BT3 9DT
Email:	h.zhou@qub.ac.uk
Tel:	+44 28(0) 9097 1753
Web:	www.ecit.qub.ac.uk/Aboutus/BusinessCard/?name=H.Zhou