In this talk, Belbachir will present background information on bio-inspired vision with examples of his research in tackling artificial vision in 2D, 3D, 4D and 3D 360° panoramic vision with few application scenarios.
State-of-the-art computer vision is narrowed into processing sequences of digital intensity images, which are synchronously produced by conventional imaging sensors, leading to highly redundant image data and inefficient processing. Biology taught us that biological visual sensing is sensitive to photometric changes and is asynchronously producing events in continuous time. With high spatial and temporal redundancy reduction, the biologically information processing system is very efficient. Conventional computer systems and biological information processing systems have a completely different architecture: computers use single very performant processing elements whose operations are sequenced in time to obtain a result; biological neural networks contain a high number of simple neurons (processors), interconnected with variable strength connections, that obtain their performance from a massive parallelism in the operations. The success of these architectures is clearly different for several types of tasks or application fields and this difference in effectiveness is probably linked to the difference in architecture and operation principle. The pre-eminence of biological systems in routine functions such as visual perception and motor control however is undisputed.
About Ahmed Nabil BelbachirAhmed Nabil Belbachir, research director at NORCE and the Head of Smart Instrumentation and Industrial Testing unit. He holds a Ph.D. in computer science from Vienna University of Technology in 2005. His specialized professional competence includes artificial vision, bio-inspired event-driven vision, 3D vision and smart cameras with a special focus on system and method design. Previously, he was involved in the ESA-Herschel project, where he was responsible for the Austrian contribution in the development of the compression software to PACS instrument from the European Space Agency. He is coordinator of few Norwegian projects on vision-based and AI-driven remote inspection. He was coordinator as well as acting as principal investigator of few European projects. He has published more than 80 scientific publications including two books: smart cameras as editor (in English 2009, in Chinese 2014) and on-board processing for infrared observatories as author in 2008. He is member of the IEEE, CLAIRE.AI, euRobotics and the general chair of the IEEE Embedded Computer Vision Workshop of 2010, 2011, 2012 and 2020. He holds 4 patents.