Identification of Beehive Piping Signals

D. Fourer and A. Orlorwska, DETECTION AND IDENTIFICATION OF BEEHIVE PIPING AUDIO SIGNALS, Proc. DCASE 2022
https://fourer.fr/dcase22

Requirements:
Dataset : https://ieee-dataport.org/documents/identification-beehive-piping-audio-signals (DOI: 10.21227/53mq-g936)

Matlab:
MSVMpack : https://members.loria.fr/FLauer/files/MSVMpack/MSVMpack.html

MFCC Implementation : https://www.jyu.fi/hytk/fi/laitokset/mutku/en/research/materials/mirtoolbox

Python:
Keras 2.2.4
Tensorflow 1.12
numpy 1.16
h5py 2.9.0

Usage:

Fig1 : generate the fig1 of the paper
start_prepare_dataset : prepare the chunks and precompute the descriptors
dataset_desc : paratition the dataset for for 3-fold cross validation

start_classif : Main script for evaluating the 4 methods 1D-CNN, TTB+SVM, MFCC+CNN and STFT+CNN through 3-fold cross-validation