DART: A Detection Audiovisual Review Tool Supporting AI Implementation in Bioacoustics
Passive acoustic monitoring enables long-term non-invasive data collection on the presence, movement, and behavior of vocal animals. However, manual analyses of this data are tedious and time-consuming. While artificial intelligence (AI) has shown great promise for augmenting or supplanting analyses...
| Main Authors: | , , |
|---|---|
| Format: | Text |
| Language: | unknown |
| Published: |
AIS Electronic Library (AISeL)
2024
|
| Subjects: | |
| Online Access: | https://aisel.aisnet.org/amcis2024/sig_hci/sig_hci/13 https://aisel.aisnet.org/context/amcis2024/article/1245/viewcontent/1515_doc.pdf |
| Summary: | Passive acoustic monitoring enables long-term non-invasive data collection on the presence, movement, and behavior of vocal animals. However, manual analyses of this data are tedious and time-consuming. While artificial intelligence (AI) has shown great promise for augmenting or supplanting analyses of acoustical datasets, validation of predictions on novel data remains a poorly supported process. Towards this challenge and to contribute design knowledge on this class of artifacts, we perform design science research to identify the objectives and requirements of this class of artifact, leading to the development of DART: the Detection Audio-Visual Review Tool. We operationalize DART in a passive acoustics study of the endangered Cook Inlet beluga whales (Delphinapterus leucas) in Alaska, USA, and evaluate the design of our artifact, demonstrating its utility and enabling faster, more reliable review of AI-processed data. |
|---|