Funded PhD Opportunity Automated approaches to feature detection and classification in benthic imagery
This opportunity is now closed.
Much of our current understanding about the distribution and ecological function of organisms and communities in the marine environment is based on analysis of imagery (Eleftheriou, 2013). Marine imaging techniques cover a wide range of scales and resolutions, from kilometres with acoustic imaging, to sub-centimetric with optical methods (Smith & Rumohr, 2013). In all respects, the number of observations decreases markedly with increasing water depth and distance from shore, meaning that for large areas of the seafloor we have very little information about the nature of the benthic environment (Webb et al., 2010). This lack of baseline data severely limits our ability to sustainably manage natural resources, understand the impacts of our activities and plan for future environmental change (Brierley & Kingsford, 2009).
The development and role of optical imagery in understanding biological processes in the marine environment has been well documented in the scientific literature (Solan et al., 2003; Durden et al., 2016; Shoening et al, 2017). Optical imagery of the benthos (photographs and video) are typically used to collect baseline information and for monitoring communities and fisheries habitats in a non-invasive manner.
These images are often used to identify organisms and characterise substrata in a fairly subjective manner that requires significant processing time. In the last decade, developments in acquisition platforms, imaging technologies and signal processing have transformed the quality and volume of data we now have available, thus motivating the current focus and development of new processing methodologies (eg. Durden et al., 2016; 2017; Gomes-Pereira et al., 2016; Schoening et al., 2017a). These developments represent significant research interest for both static imagery and video data and a wide range of potential applications (Matabos et al., 2017; Wäldchen and Mäder, 2018; Tabak et al., 2018; Schoening et al., 2017b).
This recent research paves the way to introduce computer vision and machine learning for next generation environmental monitoring which will form the basis of this PhD proposal. The ability to develop a framework that could automate, or semi-automate the process of feature detection and extraction combined with image classification in a standardised manner would be a significant advance to the field of marine ecology. Hence, the aim of this project is to explore approaches for automated classification of benthic images and to examine the ecological validity of machine learning approaches to classification.
This PhD project involves researchers from the School of Geography and Environmental Sciences and the Intelligent Systems Research Centre at Ulster University, together with external partners at the Scottish Association for Marine Sciences and the University of Limerick. It will build on a framework of existing work performed in all research centres with a focus on feature extraction and classification for benthic imagery. Classification of such images is vital for biodiversity maintenance, environmental monitoring and fisheries management and aligns closely with Ulster University’s strategic research themes of sustainability and social renewal.
Brierley & Kingsford, 2009. DOI: 10.1016/j.cub.2009.05.046 | Durden et al, 2016. Perspectives in visual imaging for marine biology and ecology: from acquisition to understanding. Oceanography and Marine Biology: An Annual Review, 54:1-72 | Durden et al., 2017. DOI: 10.1002/lob.10213 | Eleftheriou (Ed.), 2013. DOI: 10.1002/9781118542392 | Solan et al., 2003. DOI: 10.1016/S0022-0981(02)00535-X | Matabos et al., 2017. DOI: 10.1111/2041-210X.12746 | Smith & Rumohr, 2013. DOI: 10.1002/9781118542392 | Schoening et al., 2017a. 10.3389/fmars.2016.00059 | Schoening et al., 2017b. DOI: 10.1038/s41598-017-13335-x | Wäldchen & Mäder, 2018. DOI: 10.1111/2041-210X.13075 | Webb et al., 2010. DOI: 10.1371/journal.pone.0010223 |
- Upper Second Class Honours (2:1) Degree or equivalent from a UK institution (or overseas award deemed to be equivalent via UK NARIC)
- A comprehensive and articulate personal statement
If the University receives a large number of applicants for the project, the following desirable criteria may be applied to shortlist applicants for interview.
- First Class Honours (1st) Degree
- Masters at 65%
- Research project completion within taught Masters degree or MRES
- Experience using research methods or other approaches relevant to the subject domain
- Work experience relevant to the proposed project
- Experience of presentation of research findings
Vice Chancellors Research Scholarships (VCRS)
The scholarships will cover tuition fees and a maintenance award of £15,009 per annum for three years (subject to satisfactory academic performance). Applications are invited from UK, European Union and overseas students.
The scholarship will cover tuition fees at the Home rate and a maintenance allowance of £15,009 per annum for three years. EU applicants will only be eligible for the fees component of the studentship (no maintenance award is provided). For Non EU nationals the candidate must be "settled" in the UK.
- Computing, Engineering and the Built Environment
- Life and Health Sciences
- School of Computing, Engineering and Intelligent Systems
- School of Geography and Environmental Science
The Doctoral College at Ulster University
Launch of the Doctoral College
Current PhD researchers and an alumnus shared their experiences, career development and the social impact of their work at the launch of the Doctoral College at Ulster University.Watch Video
I would highly recommend Ulster University as you get so much support. Coleraine is a beautiful town and the people are so friendly. It was a really positive experience.
Carin Cornwall - PhD Environmental SciencesWatch Video
- Submission Deadline
- Monday 18 February 2019
- Interview Date
- w/c 18 March 2019
Our coastal and riverside campus focussing on science and health