JY Park PhD studentship in Imaging Science
Modulation Transfer Function (MTF) and Noise Power Spectrum (NPS) determination from natural scenes
Three years, full time
A full-time studentship is available for the project above, to candidates with Home fee status in the Faculty of Science and Technology starting in September 2017.
The offer is for a full Studentship – £16,000 annual stipend and fee waiver
The proposed PhD Imaging Science is in the area of imaging system evaluation. Image sharpness and resolution are prime factors to consider when evaluating system performance and visual image quality. Edge and detail reproduction are associated with these attributes. Multiple metrics have been developed to evaluate them, with the Modulation Transfer Function (MTF) being the most established. The MTF has traditionally been the most common system performance measure for optical systems and has its basis on linear system theory. Despite its wide use, it is well known among image engineers that no unique MTF exists for digital image capture systems, which typically employ adaptive, often non-linear spatial processing. Thus, the use of standard test targets (such as edge, sine-wave, radial grating charts), typically employed for MTF characterisation, does not necessarily return results representative of real-world performance. Recent research produced by our group proposed a newer approach to the texture-MTF measurement – designed to evaluate capturing systems – which substitutes relevant test charts with pictorial images*. We measured effective MTFs indicative of system characteristics and camera processes for a number of natural scenes. In parallel, we model image capture pipelines to determine how individual and cascaded processes/algorithms affect MTF as well as Noise Power Spectrum (NPS) measurement. The proposed PhD project aims to expand upon current work to establish a framework and relevant hardware and software set-up for producing MTF and NPS measurements directly from capturing natural scene information. Such a system is aimed for live testing of imaging sensors, capturing systems and relevant algorithms.
* R Branca, S Triantaphillidou, PD Burns. Texture MTF from images of natural scenes. In Proc. Electronic Imaging 2017: Image Quality and System Performance XIV, pp.113-120(8) 2017.
The PhD studentship is co-funded by Dr JY Park, a University of Westminster BSc/MSc/PhD alumnus, and the Faculty of Science and Technology. The graduate student will be a member of the Computational Vision and Imaging Technology (CVIT) research group of the Department of Computer Science and will be co-supervised by academic staff and relevant industry colleagues. CVIT’s research focuses on image and video formation, processing, analysis, visualisation, interpretation and evaluation. Activities of the group include the physical measurement of images and imaging system characteristics, computational methods for image analysis, visual appearance and interpretation, and interactive visual environments. Inter-relationships between physical measures, computational models and image perception are major group interests.
The student will take part in the University Graduate School and Faculty Doctoral Research Development Programme; in addition to these training programmes and the subject specific skills listed above, the student will gain important transferable skills (eg presentation skills, scientific writing and employability skills) to aid in future career progression.
Eligible candidates will hold at least an upper second class Honours degree and preferably a Master’s degree in Physical Science, Engineering, Imaging Science, Computer Vision or related disciplines. Candidates whose secondary level education has not been conducted in the medium of English should also demonstrate evidence of appropriate English language proficiency, normally defined as 6.5 in IELTS (with not less than 6.0 in any of the individual elements).
Successful candidates will be expected to undertake some teaching duties.
Informal project enquiries to Dr Sophie Triantaphillidou ([email protected]).
The closing date for applications is 5pm on 14 July 2017.