Preprint "IceDetectNet: A rotated object detection algorithm for classifying components of aggregated ice crystals with a multi-label classification scheme"
A new preprint from CLOUDLAB is available at the EGUsphere that introduces IceDetectNet, a neural network to classifiy not only shape/habit but also microphysical processes of ice crystals.
A new preprint is available from Huiying Zhang et al at the EGUsphere for public review. It presents IceDetectNet, a neural network for object detection respectively classification of ice crystals (rimed, aggregated, pure and more) based on holographic images. It distinguishes between ice crystal habit, shape and microphysical process, allowing the automatic classificaiton of large dataset. external pageRead more here.