S1.2 Eco-acoustics

25 Sep 2018
16:00 - 17:00
Lecture Hall 5

S1.2 Eco-acoustics

S1.2 _ Analysis of ecoacoustic recordings: detection, segmentation and classification. (Session Chairs: Jérôme Sueur and Dan Stowell)

Ecoacoustics is a newly emerged discipline that aims at tackling ecological research questions through the lens of sound analysis. Ecoacoustics covers several questions in marine, freshwater and terrestrial environments dealing with biodiversity monitoring, population ecology, community ecology and landscape ecology. One of the key approaches of ecoacoustics consists in identifying sounds of ecological importance in environmental recordings that were collected in an unattended way by automatic recorders. This search task is made difficult by the occurrence of background noise due to human activities, the co-occurrence of several sounds of interest, the degradation of the sounds of interest related to their propagation in the environment, a high-degree of variability of the sounds of interest, a large amount of data, and a lack of reference archives. Solutions including computer processes are currently in development to try to get around these difficulties. This session will be the occasion to report and share new techniques involving signal analysis, machine learning, deep learning and high dimension statistics for advances in detection, segmentation, supervised and unsupervised classification of sound events.

Keywords: ecoacoustics, sound analysis, classification

Talk 09-11

Information retrieval from marine soundscape by using machine learning-based source separation
Tzu-Hao Lin, Tomonari Akamatsu, Yu Tsao and Katsunori Fujikura

Acoustic sensor networks and machine learning: scalable ecological data to advance vidence-based conservation
Matthew McKown and David Klein

Content description of very-long-duration recordings of the environment
Michael Towsey, Aniek Roelofs, Yvonne Phillips, Anthony Truskinger and Paul Roe