Dynamic Curriculum Learning for Great Ape Detection in the Wild

Xinyu Yang*, Tilo Burghardt, Majid Mirmehdi

*Corresponding author for this work

Research output: Contribution to journalArticle (Academic Journal)peer-review

2 Citations (Scopus)

Abstract

We propose a novel end-to-end curriculum learning approach for sparsely labelled animal datasets leveraging large volumes of unlabelled data to improve supervised species detectors. We exemplify the method in detail on the task of finding great apes in camera trap footage taken in challenging real-world jungle environments. In contrast to previous semi-supervised methods, our approach adjusts learning parameters dynamically over time and gradually improves detection quality by steering training towards virtuous self-reinforcement. To achieve this, we propose integrating pseudo-labelling with curriculum learning policies and show how learning collapse can be avoided. We discuss theoretical arguments, ablations, and significant performance improvements against various state-of-the-art systems when evaluating on the Extended PanAfrican Dataset holding approx. 1.8M frames. We also demonstrate our method can outperform supervised baselines with significant margins on sparse label versions of other animal datasets such as Bees and Snapshot Serengeti. We note that performance advantages are strongest for smaller labelled ratios common in ecological applications. Finally, we show that our approach achieves competitive benchmarks for generic object detection in MS-COCO and PASCAL-VOC indicating wider applicability of the dynamic learning concepts introduced. We publish all relevant source code, network weights, and data access details for full reproducibility
Original languageEnglish
Pages (from-to)1163-1181
Number of pages19
JournalInternational Journal of Computer Vision (IJCV)
Volume131
Issue number5
Early online date16 Jan 2023
DOIs
Publication statusE-pub ahead of print - 16 Jan 2023

Bibliographical note

Funding Information:
We would like to thank the entire team of the Pan African Programme:‘The Cultured Chimpanzee’ Max-Planck-Institute () and its collaborators for allowing the use of their data for this project. Please contact the copyright holder Pan African Programme at http://panafrican.eva.mpg.de to obtain the source videos from the dataset. Particularly, we thank: H Kuehl, C Boesch, M Arandjelovic, and P Dieguez. We would also like to thank: K Zuberbuehler, K Corogenes, E Normand, V Vergnes, A Meier, J Lapuente, D Dowd, S Jones, V Leinert, EWessling, H Eshuis, K Langergraber, S Angedakin, S Marrocoli, K Dierks, T C Hicks, J Hart, K Lee, and M Murai. Thanks also to the team at https://www.chimpandsee.org . The work that allowed for the collection of the dataset was funded by the Max Planck Society, Max Planck Society Innovation Fund, and Heinz L. Krekeler. In this respect we would also like to thank: Foundation Ministre de la Recherche Scientifique, and Ministre des Eaux et Forłts in Cote d’Ivoire; Institut Congolais pour la Conservation de la Nature and Ministre de la Recherch Scientifique in DR Congo; Forestry Development Authority in Liberia; Direction des Eaux, Forłts Chasses et de la Conservation des Sols, Senegal; and Uganda National Council for Science and Technology, Uganda Wildlife Authority, National Forestry Authority in Uganda.

Funding Information:
We would like to thank the entire team of the Pan African Programme:‘The Cultured Chimpanzee’ Max-Planck-Institute (2022) and its collaborators for allowing the use of their data for this project. Please contact the copyright holder Pan African Programme at http://panafrican.eva.mpg.de to obtain the source videos from the dataset. Particularly, we thank: H Kuehl, C Boesch, M Arandjelovic, and P Dieguez. We would also like to thank: K Zuberbuehler, K Corogenes, E Normand, V Vergnes, A Meier, J Lapuente, D Dowd, S Jones, V Leinert, EWessling, H Eshuis, K Langergraber, S Angedakin, S Marrocoli, K Dierks, T C Hicks, J Hart, K Lee, and M Murai. Thanks also to the team at https://www.chimpandsee.org. The work that allowed for the collection of the dataset was funded by the Max Planck Society, Max Planck Society Innovation Fund, and Heinz L. Krekeler. In this respect we would also like to thank: Foundation Ministre de la Recherche Scientifique, and Ministre des Eaux et Forłts in Cote d’Ivoire; Institut Congolais pour la Conservation de la Nature and Ministre de la Recherch Scientifique in DR Congo; Forestry Development Authority in Liberia; Direction des Eaux, Forłts Chasses et de la Conservation des Sols, Senegal; and Uganda National Council for Science and Technology, Uganda Wildlife Authority, National Forestry Authority in Uganda.

Publisher Copyright:
© 2023, The Author(s).

Fingerprint

Dive into the research topics of 'Dynamic Curriculum Learning for Great Ape Detection in the Wild'. Together they form a unique fingerprint.

Cite this