pcbi.1009706.g004.tif (1.2 MB)
Download fileEcho classification using convolutional neural networks.
figure
posted on 2021-12-16, 18:59 authored by Ralph Simon, Karol Bakunowski, Angel Eduardo Reyes-Vasques, Marco Tschapka, Mirjam Knörnschild, Jan Steckel, Dan Stowell(a) Architecture of the neural network. (b) Results of the CNN model on two different iterations of the data and different amounts of echoes as the input. (c) The confusion matrix of a trained 10-input CNN on the testing set.
History
Usage metrics
Read the peer-reviewed publication
Keywords
usually concave shapedspectral target strengthsouth east asiaremains unclear whetherhigher classification performancegood classification accuracyconvolutional neural networkflowers acoustically conspicuousinvestigate acoustic traitsacoustic plant specializationsxlink "> bat32320 flower echoes12 different speciespollinated flowers comparedacoustic traitsflower echoesdependent plantpollinated flowersflowers belongingtwo batpollination syndromespollinated speciesmorphological featureslower variance84 %)
CC BY 4.0