Food scientist Dr Paul Williams is among 13 distinguished young professionals recognised by SPIE, the international society for optics and photonics, as one of it is 2020 DCS Rising Researchers. Dr Williams is a lecturer and researcher in the Department of Food Science in Stellenbosch University's Faculty of AgriSciences.
He is the only academic among the thirteen in the 2020 DCS Rising Researcher group not associated with an American university. They will be recognised on 27 April 2020 at the SPIE Defence + Commercial Sensing (DCS) 2020 conference in California in the USA.
According to a press release by SPIE, its Rising Researcher initiative is now in its fourth year. It recognises early-career professionals who are conducting outstanding work in product development or research in the areas of defence, commercial, and scientific sensing, imaging, optics, or related fields such as astronomy and food science.
At SPIE DCS 2020, he will present some preliminary findings on the use of near infrared (NIR) hyperspectral imaging to distinguish between different types of game meat and game meat cuts. This work, together with his postgraduate students on its use as an ID method for South African game species such as springbok and blesbok, follows on those done by other researchers on kangaroo and reindeer meat.
Dr Williams says he received the good news about his inclusion late December whilst on honeymoon. At the time, he was checking emails because an urgent funding application was due. “Funders do not care about honeymoons," he dropped.
His field is all about how light interacts with a product, and how the highly sensitive camera within imaging instruments picks up on the different chemical signals and converts it into images. “With NIR we can see chemical differences or similarities on the computer that we cannot see with our eyes. It allows us to visualise the potential differences or similarities between different materials or objects, from food to minerals. Each have their own chemical 'signal'," he explains.
Dr Williams says the issue of food fraud was top of mind when he started this work. “Once it is cut, you cannot really distinguish kudu from springbok meat, for instance, and therefore we need methods that can do so."
Current DNA identification methods can be quite costly, and time-consuming.
“Our NIR studies have already shown that there are definite chemical differences between the meat of species, such as blesbok and springbok and between different muscles in their bodies," he explains.
Dr Williams specializes in the use of vibrational spectroscopy and imaging techniques, such as NIR spectroscopy and hyperspectral and multispectral imaging. He is also interested in data analysis. To this end he has this year returned to student life. He is part of the first intake of students to follow a new postgraduate industrial engineering diploma with a focus on data science at Stellenbosch University.
Dr Williams has been using imaging techniques in his research since his masters degree years under supervision of now colleague Prof Marena Manley of the Department of Food Science.
One word was all it took to interest him in the field. “While I once was chatting with Prof Manley about possible postgraduate options, she mentioned the word 'chemometrics'. It intrigued me, although I did not know what it meant," he remembers how he became interested in the science of extracting information from chemical systems by data-driven means.
When he googled chemometrics he found that it involves a lot of maths, statistics and potentially also computer programming.
“I just knew I wanted to do that," emphasises Williams, who over the years has taught himself some coding.
In 2007, the year that he started his masters' degree studies in food science, he attended his first conference on NIR in Sweden – and was hooked for life.
He used this technique during his PhD studies to work on a non-destructive way to grade and evaluate the quality and safety of maize kernels.
A doctoral student, Kate Sendin, has since built upon his initial work, to develop what could be a “lite" version of a NIR hyperspectral imaging system to easily do relevant testing on site. “The next step would be to involve engineers and designers to develop a cheaper, more specific system that can be commercialised and used in industry by silo managers and the likes," he says.
- To date, Dr Williams has published 19 papers in peer-reviewed journals and has supervised, to completion, 2 PhD students and 10 MSc students.