A4 Article in conference proceedings
Improving Taxonomic Image-based Out-of-distribution Detection With DNA Barcodes (2024)
Impiö, M., & Raitoharju, J. (2024). Improving Taxonomic Image-based Out-of-distribution Detection With DNA Barcodes. In 32nd European Signal Processing Conference (EUSIPCO 2024) : Proceedings (pp. 1272-1276). IEEE. European Signal Processing Conference. https://doi.org/10.23919/eusipco63174.2024.10715139
JYU authors or editors
Publication details
All authors or editors: Impiö, Mikko; Raitoharju, Jenni
Parent publication: 32nd European Signal Processing Conference (EUSIPCO 2024) : Proceedings
Place and date of conference: Lyon, France, 26.-30.8.2024
ISBN: 979-8-3315-1977-3
eISBN: 978-9-4645-9361-7
Journal or series: European Signal Processing Conference
ISSN: 2219-5491
eISSN: 2076-1465
Publication year: 2024
Publication date: 26/08/2024
Pages range: 1272-1276
Number of pages in the book: 2761
Publisher: IEEE
Publication country: United States
Publication language: English
DOI: https://doi.org/10.23919/eusipco63174.2024.10715139
Publication open access: Not open
Publication channel open access:
Publication is parallel published (JYX): https://jyx.jyu.fi/handle/123456789/99083
Web address of parallel published publication (pre-print): https://arxiv.org/abs/2406.18999
Abstract
Image-based species identification could help scaling biodiversity monitoring to a global scale. Many challenges still need to be solved in order to implement these systems in real-world applications. A reliable image-based monitoring system must detect out-of-distribution (OOD) classes it has not been presented before. This is challenging especially with fine-grained classes. Emerging environmental monitoring techniques, DNA metabarcoding and eDNA, can help by providing information on OOD classes that are present in a sample. In this paper, we study if DNA barcodes can also support in finding the outlier images based on the outlier DNA sequence's similarity to the seen classes. We propose a re-ordering approach that can be easily applied on any pre-trained models and existing OOD detection methods. We experimentally show that the proposed approach improves taxonomic OOD detection compared to all common baselines. We also show that the method works thanks to a correlation between visual similarity and DNA barcode proximity. The code and data are available at https://github.com/lmikkoim/ldnaimg-ood.
Keywords: biodiversity; monitoring; determination of species; classification; systematics; computer vision; DNA barcodes; signal processing; machine learning
Free keywords: image-based taxonomic identification; out-of-distribution detection; DNA barcodes
Contributing organizations
Ministry reporting: Yes
VIRTA submission year: 2024
Preliminary JUFO rating: 1