Top software students tackle current challenges
An exceptional student from the University of Dundee has been named Young Software Engineer of the Year 2018. Can Gafuroglu won first prize for an innovative project using AI and machine learning to support smart diagnosis of early dementia.
Second place was won by Radostin Stoyanov from the University of Aberdeen for his work on the live migration of Linux containers. Nikita Samarin from the University of Edinburgh won third place with an innovative project investigating the use of an ECG based biometric for personal authentication.
The Leidos Software Engineering Project Award went to Jonathan Robson from the University of Abertay for developing a visibility buffer rendering system for use in the games sector, particularly for VR headsets.
The Young Software Engineer of the Year Awards, now in their 29th year, are given to the best undergraduate software projects, drawn from across all students studying computing science and software engineering in Scotland. Each university submits the best final year undergraduate software engineering project from amongst their students.
The Awards are organised by ScotlandIS, the trade body for the digital technologies industry and were presented at the ScotSoft2018 dinner, with over 550 guests from across the industry.
Brilliant young minds making breakthroughs in computing science and programming:
University of Dundee
“Joint prediction and Classification of Brain Image Evolution Trajectories from Baseline with Application to Early Dementia Diagnosis”
Can developed a strong interest in the increasingly important field of machine learning during his third year at the University of Dundee and has used that to create a predictive model that supports the early diagnosis of dementia from limited data.
There have been a number of research studies into the onset of dementia, but they are often dependent on multiple data points which are difficult to collect, as they involve checking the patient over a period of time to analyse the subtle brain changes that occur as time progresses. Health professionals have to balance the time delays involved in collecting this greater level of data with the need for early diagnosis and treatment.
Can’s project uses machine learning to extrapolate and enhance more limited observations. His project brings together a range of existing datasets both longitudinal and cross-sectional to enhance the ‘learning’ from individual observations. The program learns how to predict the developmental trajectory of a brain from a single acquisition timepoint, classifying the predicted trajectory as either ‘healthy’ or ‘demented’. Using a weighted system to verify the classification Can’s programme improved overall performance and analysis by 5%.
Can is originally from Turkey but was educated in St Andrews, before studying Computer Science at University of Dundee. He is now working on a games development project. His outside interests include travelling, learning languages, playing the guitar, and good food.
The judges considered this project to be exceptional. As overall winner of the Young Software Engineer of the Year Award Can received a cheque for £2500 from Sopra Steria, and a trophy from ScotlandIS.
University of Aberdeen
“Efficient Live Migration of Linux Containers”
With the increasing use of cloud computing, one of the challenges facing developers and IT professionals is the need to migrate data, processes and applications from ‘on premise’ to the cloud. This needs to be done in an efficient way that minimises use of hardware resources and enables no apparent lack of continuity during the migration. This can be particularly challenging where applications are running modifications faster than they can be transferred over the network to the cloud. “Containers” have been developed to ensure efficient and effective migrations of data and processes in a cloud environment and is a widely used technique. However, in certain technology environments ‘containers’ can still prove problematic.
Linux is one of the building blocks of the software industry but migration of Linux containers has proved challenging, consuming significant time and resources, and impacting performance.
Radostin’s project set out to solve this problem enabling the simplification of live Linux container migration using an image proxy based transfer mechanism. Having developed the system he then evaluated it against existing alternatives, and found his approach delivered greater efficiency, and resulted in lower total migration times and application downtime, reducing the complexity of the migration process.
During the development of his project Radostin contacted community groups specialising in containers. They are proposing to use the results of his project in future cloud computing applications – a great endorsement of his work!
The judges were impressed by this smart solution to a complex current technology challenge and considered Radostin to be a well deserved runner-up. He received a cheque for £2000 from sponsor BCS, and a trophy from ScotlandIS.
University of Edinburgh
“A Key to Your Heart: Biometric Authentication Based on ECG Signals”
Another international student, Nikita hails from Estonia and came to Scotland to study Computer Science at the University of Edinburgh.
His project investigates the growing field of biometric authentication. As we make use of biometrics for increasing range of applications, from unlocking our smartphones, to accessing banking services and and augmenting our passports, it is important to ensure they provide the necessary level of security, robustness and usability.
Intrigued by this problem, Nikita harnessed evidence from earlier research that demonstrated that electrocardiogram (ECG) signals are sufficiently unique to each individual to be suitable for user authentication. He set out to design and implement an ECG based biometric authentication system that was accurate, stable and non-intrusive.
As part of his project he collected data from a test group using a consumer grade ECG monitor and ran a series of experiments to verify whether ECG biometrics would be sufficiently robust to be used for personal authentication. The results were very encouraging. So if next time you log into your bank account and it asks for an ECG reading you heard it here first..
Nikita won third prize for his well executed project, a cheque for £1500 from sponsor Edge Testing and a trophy from ScotlandIS.
Having worked in a number of technology companies as a placement student during his undergraduate course and spent an exchange year with the University of California, Nikita has decided to pursue his research interests and is about to start a PhD at the University of California, Berkeley
“An Evaluation of Applications of Visibility Buffer Rendering Techniques in Transparent Geometry Rendering”
Real-time rendering of computer graphics, especially video game graphics, is a field that is constantly advancing to take advantage of new developments in computer hardware. With this increasingly powerful hardware, there is an expectation amongst consumers for higher quality video games graphics. The advent of high resolution displays – for example 4K monitors and TVs and virtual reality (VR) – is driving demand for higher resolution rendering.
Jonathan sought to tackle some of the issues surrounding higher resolution rendering, by creating an algorithm using a new technique, called transparent rendering, to speed up the rendering process and improve the quality of the user experience.
The new algorithm proved very successful in low complexity scenes but highlighted the need for further development in fast moving environments.
The judges were impressed by the quality of this project and its adherence to best practice software engineering principles, and awarded Jonathan the Leidos Software Engineering Award.
In addition to the Award, Jonathan received a cheque for £1500 donated by Leidos, and the Leidos Trophy was given to his university, and the University of Abertay.