A new research chair in computer vision at ÉTS

École de technologie supérieure (ÉTS), in collaboration with Matrox Electronic Systems Ltd, recently created the Matrox Imaging Industrial Research Chair in Computer Vision for Industrial Applications to help companies forge their own space within the global computer vision market.

According to Christian Desrosiers, holder of the new Chair and Co-Director of the Imaging, Vision and Artificial Intelligence Laboratory (LIVIA) at ÉTS: “The objective of our research is to design specialized algorithms in the field of computer vision that will perform well even with a limited quantity of annotated data. We are also aiming to ensure that these algorithms can adapt quickly to various environments and conditions.”

It is worth noting that, considering escalating demand in the areas of automation, inspection, and connected objects, the international market for computer vision is expected to growth at an annual rate of 26.3% by 2026. Computer vision will also be called upon to play a major role in the field of medical imaging.

However, improvements must be made to learning techniques based on deep neural networks, otherwise known as deep learning, in order for computer vision to attain its full potential and to facilitate the transfer from research to industry.

These improvements are required on multiple fronts, including mass data-collection and labelling. These two tasks, often carried out manually, require considerable effort on the part of the teams involved. The collection of labelled images is particularly complex for under-represented categories of data, such as anomalies, which are crucial to the design of inspection systems.