Pollen grain recognition through deep learning convolutional neural networks

Palynology is the study of pollen, in particular, pollen’s grain type, but tasks classification and counting pollen grains are highly skilled laborious. Despite efforts made during last decades, manual process still predominant. One reasons for that small number taxa usually used previous approaches. In this paper, we propose a new method to automatically classify using state-of-the-art deep learning technique applied recently published POLEN73S image dataset. Our proposal manages up 94% samples from dataset with 73 different classes grains. This result, which surpasses all attempts difficulty under consideration, gives good perspectives achieve perfect score recognition task even large types.