Computer Vision News - December 2020

3 Erdem V. and Amin N. 19 would need to go through each of the images or timeframes in a video, find the neurons and extract signals from them. Now, it is an automated process . They can just input the videos and grab the neural activities from them, which means experiments can be conducted in a high-throughput manner, leading to greater scientific insight. “What I am most proud of about this work is not only does it have a great deal of potential in terms of accelerating the scientific hypothesis testing cycle, which is of great use to neuroscientists, but it can be of high interest to computer vision practitioners and method developer s as well,” Erdem enthuses. “The computational techniques we put forward could inspire other computational techniques for a completely different set of problems.” Erdem and Amin have made all the codes and software available online . They would like to encourage people whoworkwithC. elegans or othermodel organisms to explore their papers and see if they can use anything from the pipeline in their own problems. Best of MICCAI 2020