At Amsterdam Science Park, we believe knowledge has no borders. That’s why you’ll find cutting-edge institutes, top-class facilities and business incubators in one convenient location. You’ll meet motivated students, innovative researchers, talented scientists and adventurous entrepreneurs – all in an environment where it’s easy to work together. Cross-disciplinary initiatives and holistic thinking flourish. Ground-breaking research projects become successful spin-offs. That’s the strength of Amsterdam Science Park: connecting boundless minds.
Cees Snoek is a video and image retrieval expert involved in both research and high-tech industry at Amsterdam Science Park.
“For me, it all started back in the 1990s when I was doing a Master’s in Business Information Systems at the University of Amsterdam (UvA). I happened to attend a lecture about computer algorithms that could recognise faces in images and that was it – I was hooked! Subsequently I specialised in image recognition for video search and I’ve pursued image recognition every since, gaining my PhD at UvA and doing research at UC Berkeley and Carnegie Mellon University."
“I see myself as a researcher who is involved in high-tech industry – it’s the best of both worlds”
“My professor at UvA started Euvision Technologies, a spin-off company, in 2010. I became its Head of Research and Development. We saw a business opportunity and we were ahead of the crowd: it took a while for the rest of the world to catch up. We specialised in image recognition applications powered by artificial intelligence. Our first publicly released app was Impala — an app for iOS and Android that essentially ‘reads’ and organises photos on your phone into different categories, without needing to work in the cloud."
“In 2014, the research arm of Qualcomm, the American multinational semiconductor company, acquired Euvision, and so we became Qualcomm Research Netherlands. I am head of research here and I still teach at UvA, where I supervise seven PhD students and two post-doctoral researchers. I am also one of the three directors of QUVA lab, which is a world-class research institute created in 2015 by the public-private partnership of UvA and Qualcomm. The core business of QUVA Lab is computer vision and deep learning, especially where the two meet (which we call deep vision). In addition, I’m involved in the national ICT research programme, COMMIT. I see myself as a researcher who is involved in high-tech industry – it’s the best of both worlds."
“Amsterdam Science Park is the best place to be in the Netherlands for image recognition research, and it has a lot to offer as a base”
“When I started with image recognition research, we needed a different algorithm to identify every different category within an image. Then the paradigm changed, and a single algorithm could learn to recognise different things – a major shift which boosted the technology enormously. UvA was very involved in this rapid development. Now the technology has matured to reach the large-scale consumer market, with the likes of Facebook’s face recognition software."
“The next frontier, which is very relevant for the development of the self-driving car, is the recognition of moving objects and people, which is still very hard to do. Moving image recognition would also be very helpful for security, by identifying potential threats. We’re also working on very finely grained identification – being able to distinguish between 200 species of birds, for example. Additionally, I’m involved in trying to achieve a precise translation of video into text – there is just too much video to watch now, so imagine if we could automatically summarise each film into a paragraph of text, flagging anything unusual or interesting."
“One reason we’re here is the presence of talented people with the right skills”
“Amsterdam Science Park is the best place to be in the Netherlands for image recognition research, and it has a lot to offer as a base. Back in 2000, I must admit I found it quite boring here, but since the Faculty of Science was established and we acquired a train station, I’ve seen an enormous improvement. It’s now a lively hub of science and innovation. At Qualcomm Research we currently employ eight people, and one reason we’re here is the presence of talented people with the right skills in computer vision and machine learning. When I graduated, there weren’t many options in image recognition, but now opportunities are increasing and I would recommend it as a career direction."
“I do think the park can improve. While the presence of so many institutes at the park is a big plus, presence is not enough. There needs to be interaction, but being situated close together is an important first step.
I would like to see more opportunities for matchmaking between scientists with an innovative idea and entrepreneurs who know how to run a company. Cross-fertilisation is beneficial and inspirational for all concerned."
“I would like to see more opportunities for matchmaking between scientists and entrepreneurs”
“In QUVA lab, we have about 20 people, and while industry provides inspiration in the form of problems, we give these into a scientific angle It’s use-inspired research. In the future, the world will be making more and more use of visual data. The Internet of Things may mean that all smart devices contain a camera – meaning a lot of potential for behavioural research. Privacy is the big issue here, but it’s a development that could give us interesting scientific insights. Searching large broadcast-recording archives is another application. All in all, the question of how to search and make use of the huge amount of images surrounding us is one that will be here for a while, and we intend to rise to that challenge at Amsterdam Science Park.”