Sorry about obvious question, but with recent advances in computer vision/artificial neural networks (CV/ANN) are there any applications of it to mozak purposes?
Did anybody tried to set up ANN to do the tracing job? Or mozak’s goal is to get initial labelled set for future application of ANN?
using automated techniques to reconstruct neuron images is an active area of research at the Allen Institute. working on Mozak, we don’t do much day-to-day with machine learning or computer vision techniques for auto-reconstruction, and creating labeled sets for machine learning isn’t necessarily an explicit research goal, but enabling other advances in automatic reconstruction – especially using machine learning techniques, which are very data-hungry – is definitely a thing that’s at the forefront of our mind.
thanks for the question!
The more data we validate, the better the training data for future ML.