At NR we have over 30 years of experience working in image analysis and machine learning and our methods are used in a wide range of industries, such as healthcare, transport, ocean, climate and environment and technology.
We work with various types of image data from cameras and sensors, and extraction, characterisation and object recognition of images and image sequences are central themes for many of our projects. We also have experience with other types of data, like acoustic signals, audio recordings and text.
In healthcare, we are developing methods to recognise signs of disease on x-rays and ultrasounds, while we work closely with partners in the marine industry to calculate fish stocks using underwater acoustics. In the IARI project, NR is collaborating with Bane Nor and developing algorithms that recognise faults on the trainlines infrastructure by analysing images taken by cameras on trains and drones.
Regardless of the application, efficient solutions are usually found with methodological approaches that combine prior knowledge, context and observed data.
Main research areas
We use deep learning methodologies to collect, detect and classify large volumes of observational data from the marine industry. This can be sonar acoustics, images and videos captured from trawls or the seabed, drone imagery of marine mammals, or different types of microscopic imagery. The data contains invaluable information in order to monitor marine stocks and ecosystems.