Computer Vision Workshop
Discussions took place formally in the auditorium and continued more informally during the pizza evening and other side events. The participants also visited four of the CAnMove facilities that uses computer vision methods (MagnetoreceptionLab, Wind tunnel, NanoLAB, LUMBO), participated in the thermal cameras demo by Axis Communication and visit the IKDC 3D-printing facility at Lund University Design Center.
We were particularly pleased to see 10 submissions to the Animation Competition. The quality of the animations was so high that has been difficult to decide which was the best. The prize was awarded to Atticus Pinzón-Rodríguez that presented an animation with the title “Tracking the orientation behavior of zebra finches under controlled light and magnetic conditions”.
We are facing a fast technological development in terms of 3D shape analysis and many of the new techniques have huge potentials in animal studies. 3D shape technologies are already used to study the shape of deforming wings in birds and bats (Christoffer Johansson), to the study of phenotypic variation (Erik Svensson) or animal brain morphology (Alexander Kotrschal). However, 3D-scanning techniques offer great opportunities, but also many challenges in terms of post processing and data analysis (Stefan Lindgren).
Tracking is one of the most common application of computer vision and widely used in animal behavior studies. However, tracking is never a trivial task especially when diverse animals are studied (Rachel Muheim). Complexity in tracking algorithms need state of the art calibration techniques (Tobias Palmér), the use of appropriate motion models (Clas Veibäck) and even newly implemented sensors setups (Samuel Jansson).
We learnt about the most recent achievements in the field of computer vision and how such new approaches will be used in the future in farming (Mikael Nilsson) or clinical analysis (Rasmus R. Paulsen). However, the field of image analysis is constantly improving opening up new possibilities (Michael Felsberg) that can also be integrated, and sometimes replaced, by different tools including radio and audio signals (Kalle Åström).
Finally, during the open discussion, we tried to identified the mutual interests among researchers in computer vision and animal behavior and morphology. We agreed on the future steps to take to facilitate the collaboration between departments and universities and investigated grant applications strategies to build on fruitful collaboration.
We would like to thank all the participants of the workshop for their active contribution and all delegates who helped us during the event. See you at the next meeting with new interesting collaborative results!
/The organization committee,
Christoffer Johansson, Giuseppe Bianco and Christina Rengefors
Text: Giuseppe Bianco/Photo: Sylvie Tesson