Developing a revolutionary breakthrough in road surveys

Company

Dr Stephen Remde, Director, Innovation and Research 

Gaist has always been ambitious and aims to maintain a leading position with the effective use of new technologies and innovation. The first KTP project between Gaist and the University of York has rapidly built Gaist's reputation in the prediction and modelling of long-term deterioration of infrastructure through the use of artificial intelligence, modelling and optimisation. With the outcome of the second KTP project, Gaist sees a huge potential to exploit the advanced knowledge of Computer Vision and Machine Learning to deliver more robust and cost-effective autonomous solutions in condition surveys on road surfaces, pavement and infrastructure as well as asset management.

Recently this year, Gaist secured £2.7 million of investment from BGF to further aid in its growth into other domestic markets and overseas, and to invest into its ground-breaking computer vision products. Gaist have established a new computer vision and machine learning research team and recruited more scientists and researchers to join the company. With our massive collection of surveying data together with the growing demand for large-scale data analysis, we aim to hold our position as the leading company for providing road intelligence data.

KTP Associate

Dr Supannee Tanathong, Computer Vision Scientist

The KTP project was an ideal preparation for me to enter into the commercial workplace while still receiving supervision from my academic advisor and being able to access academic resources.

Being involved in this KTP project provided me with the unique opportunity to work in both an industrial environment and academia which helped me decide which types of jobs I would like to pursue. I found I prefer working within a company and I continued working at Gaist as a Computer Vision Scientist after my KTP project ended.

During the KTP, I had been involved in every part of the project from measuring the field of view of the cameras during fitting, calibrating them, and developing several computer vision software from small tools to commercial products as well as writing a patent document for our innovation. I had enhanced my skills in computer vision and machine learning through the supervision of my KTP academic advisor, Dr. William Smith. I had full access to the company dataset, their survey tools, as well as their web infrastructure which made it more convenient for me to build my own software and visualise my research outcomes. I also took advantage of various training courses and attended several conferences using the KTP development budget.

Academic

Dr William Smith, Senior Lecturer in Computer Vision

From an academic perspective, the KTP project was both a very productive and eye opening experience. Productive because Gaist are a very forward looking company, happy to take on risky and challenging research problems. Eye opening because it brought me face to face with real world problems and data of a scale I had never contemplated before! I can genuinely say that I approach all my research with a different perspective having now worked with datasets comprising millions of images captured in highly variable real world conditions. I learnt many new skills, gained an understanding of how research feeds into products and uncovered many further research questions that may lead to future projects or publications. In concrete terms, I gained my first experience of protecting intellectual property as we successfully patented part of our work. We published a workshop paper and have a paper at a top journal under review. The research from the project contributed directly to a new product and I was able to set masters and undergraduate projects related to the work done on the KTP. I now have a PhD student working on follow up research questions. Overall, I was pleasantly surprised by how successful the KTP formula is. If you find the right company with the right approach, it provides an opportunity to do applied research with real and immediate impact but also exposes you to new problems and data that can lead to fundamental research questions.