Computer Vision News - March 2023

42 WASPSYN Challenge at ISBI neural networks, artificially generated or manually designed networks, but we’re studying real-world biological neural networks. ” The brain is represented as an extensive graph, with neurons as nodes connected by synapses . Every synapse is an edge in this directed graph, and there can be multiple edges in these two nodes because there can be multiple synapses connecting two neurons. “ Theneuronssplitouttheneurotransmitters The objective of the WASPSYN Challenge , part of the ISBI 2023 conference in April, is to detect synapses within electron microscope images . It is an out-of-domain challenge, introducing three brain samples, with participants expected to train on one and test on the other two. The task represents a significant challenge in connectomics to understand the intricate network of neuron connections within the brain. The brain is sliced into thin sections – in this case, a mere 8 nanometers (nm). The resulting voxel size is 8 x 8 x 8 nm, providing extremely high- resolution 3D imagery. “ The goal is to understand the complete biological network in the brain,” Jingpeng tells us. “We normally talk about artificial Jingpeng Wu is an Associate Research Scientist at the Flatiron Institute in New York. He is organizing an interesting ISBI challenge exploring the potential of out-of-domain generalization methods for large-scale connectomics applications. He speaks to us about what is involved. WASPSYN CHALLENGE AT ISBI 2023