BatNet++: A Robust Deep Learning-Based Predicting Models for Calls Recognition

Biodiversity is the most complex feature of our planet and it is the most vital. Without biodiversity, there is no future for humanity. People are gradually realizing the significance of maintaining biodiversity. Since the lack of effective biodiversity detection tool, human perform inefficiency in biological detection and protection. In this paper we proposed a novel method, called BatNet++, in order to monitor biodiversity of bats by identifying bat calls. Experimental results show that BatNet++ yields effective improvement in bat calls recognition. A large amount of experiments shows that our method can be harnessed for given taxa which make consistent sounds that may be used as vocal barcodes to identify species (e.g., amphibians and birds).