Project Application Details

ARCHIE Project title: 30
Project PI: School of Engineering and Computing
Department: 01 September 2014
Address: School of Engineering and Computing, , licensed software, PA12BE
Relevant industrial collaborations:
Existing relevant KTP activities:
Does the PI require to run calculations? 8
Number of users involved in the project: 3 years
User 1 details:

Ahmad Alzubi
Email:Ahmad.Alzubi@uws.ac.uk
School of Engineering and Computing
University of the West of Scotland

User 2 details:

Meshal Alqahtani
Email:B00262120@studentmail.uws.ac.uk
School of Engineering and Computing
University of the West of Scotland

Service level requested:
ARCHIE Project description: 30
Potential output of the project:

Within the context of media search enabling technologies, research into general-purpose multimedia analysis and its efficient retrieval has gained significant momentum. In addition, the size of Web is growing rapidly with different forms of data (i.e. text, images, audio, and videos), and due to the huge development of visual contents, the demand for video services (e.g. watching, downloading, and sharing) has become more imperative. Most of popular video search engines provide video retrieval services using text-based approach, and facilitating self-broadcasting option which causes problems associated with inaccuracies.
This project aims at developing novel information retrieval techniques for media search focusing on search engines for images and video data.
The objectives of this research can be broadly summarised as follows:
• Investigating new software hardware co-design approaches for image and video data retrieval using latent semantic indexing (LSI);
• Accelerating the retrieval of media data using hardware accelerators and infrastructure in the form of graphical processor unit (GPUs), cloud computing and High Performance Computing (HPC) platforms;
• Deploying appropriate analysis techniques and performance measures in video quality enhancement using LSI; and
• Handle the semantic gap problem between low-level features extracted from video and the user’s need, and meaningfully interact with a higher level features i.e. semantic.
The proposed research will be carried out as follows:
• Work package 1 (3 months):
1. Carrying out literature review on different techniques used in video retrieval systems
2. Compiling the system specifications and requirements
• Work package 2 (6 months):
1. Understanding the hardware accelerators and environment to be used of the investigation
2. Perceiving video retrieval and processing
• Work package 3 (6 months):
1. Designing efficient algorithms, methodologies, and models with the aim of building a video retrieval system using hardware accelerators
• Work package 4 (9 months):
1. Building the system according to the conducted algorithms and specifications.
2. Applying user queries
3. Representing the results
• Work package 5 (6 months):
1. Carrying out experiments to verify and validate the results of the proposed system
2. Analysing and evaluating the results in terms of computation time and accuracy
3. Writing reports and publications
• Work package 6 (6 months):
Thesis write up and publications

Project start date: 01 September 2014
Duration of the project: 3 years
Total CPU time required (in hours): 1000
CPU hours required by typical job: 8
Approximate amount of data created per job (in GB):
Approximate amount of data created per project: 30
Funding Agency Grant Reference No:
pFact ID:
RKES ID:
Software required for the project: licensed software
Software required to be installed for the project: Matlab
Experience in using software package 1:
Software required to be installed for the project:
Experience in using software package 2: 8
Software required to be installed for the project:
Experience in using software package 3: 8
Training required:
Other training required:
Experience in using HPC: 5