DeepForestVisionV2: Ecology-Driven Taxonomy Expansion for Camera-Trap Monitoring in African Tropical Forests
Abstract
Camera-trap monitoring in African tropical forests increasingly extends beyond closed-canopy interiors to riverbanks, clearings, and park edges. Among available open tools for African forest camera-trap classification, DeepForestVision is the only one providing a matched offline workflow for both photographs and videos, and previous work showed that it outperformed other available baselines on a comparable benchmark. However, it was designed for closed-canopy, ground-level forest interiors and uses a 35-class prediction space that becomes too coarse when deployments encounter arboreal primates, birds, semi-aquatic taxa, or human-associated confounders such as livestock. We present DeepForestVisionV2, an ecology-driven expansion from 35 to 64 prediction classes (61 animal classes plus human, vehicle, and blank) designed to address three recurrent deployment gradients: vertical stratification, scene openness, and anthropogenic interfaces. DeepForestVisionV2 retains the same offline workflow and is trained on 1,535,010 photographs and 243,354 videos from multi-country African tropical-forest projects. Evaluation combines a cross-country cropped-photo validation set, used to assess robustness across sites and camera-trap settings, with three held-out Uganda video benchmarks spanning the targeted gradients. On the validation set, DeepForestVisionV2 reaches 0.86 accuracy, 0.82 macro-F1, and 0.81 balanced accuracy. On the deployment benchmarks, it preserves or improves baseline accuracy despite its harder classification task, while increasing the number of identified taxa from 22 to 29 in forest-interior videos and from 4 to 9 at riverbanks. In the park-edge use case, it raises accuracy from 0.62 to 0.86 and reduces false alarms from 11 to 0. These results show that DeepForestVisionV2 materially improves field utility while preserving robustness across sites, habitats, and camera-trap settings.
Versions
Related
- Climate sensitivity effects on tropical bat ecology
- Monitoring the seasonal functioning of tropical ecosystems with phenocams
- Acoustic Monitoring of Tropical Bats: Saccopteryx bilineata's response to habitat and time across a restoration gradient using a 2D-CNN
- Towards Automated Acoustic Monitoring of Threatened Birds in Tropical Forest Ecosystem Restoration
- Phylogeny is not taxonomy: lessons on higher classification from turtle and whale barnacles (Cirripedia: Coronuloidea)
BibTeX
@misc{magaldi2026deepforestvisionv2,
title = {{DeepForestVisionV2: Ecology-Driven Taxonomy Expansion for Camera-Trap Monitoring in African Tropical Forests}},
author = {Hugo Magaldi and Theau d'Audiffret and Etienne Francois Akomo-Okoue and Bala Amarasekaran and Naomi Anderson and Claire Auger and Noemie Cappelle and Daniel Cornelis and Raphael Cornette and Tobias Deschner and Gabriel Dubus and Davy Fonteyn and Rosa M. Garriga and Jennifer Hatlauf and Innocent Kasekendi and Raymond Katumba and Aram Kazandjian and Alfred Ngomanda and Stephan Ntie and Simone Pika and Xavier Rufray and Harold Rugonge and John Justice Tibesigwa and Peter van Lunteren and Hadrien Vanthomme and Joeri A. Zwerts and Sabrina Krief},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2606.20223},
eprint = {2606.20223},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2606.20223},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - DeepForestVisionV2: Ecology-Driven Taxonomy Expansion for Camera-Trap Monitoring in African Tropical Forests
AU - Hugo Magaldi
AU - Theau d'Audiffret
AU - Etienne Francois Akomo-Okoue
AU - Bala Amarasekaran
AU - Naomi Anderson
AU - Claire Auger
AU - Noemie Cappelle
AU - Daniel Cornelis
AU - Raphael Cornette
AU - Tobias Deschner
AU - Gabriel Dubus
AU - Davy Fonteyn
AU - Rosa M. Garriga
AU - Jennifer Hatlauf
AU - Innocent Kasekendi
AU - Raymond Katumba
AU - Aram Kazandjian
AU - Alfred Ngomanda
AU - Stephan Ntie
AU - Simone Pika
AU - Xavier Rufray
AU - Harold Rugonge
AU - John Justice Tibesigwa
AU - Peter van Lunteren
AU - Hadrien Vanthomme
AU - Joeri A. Zwerts
AU - Sabrina Krief
PY - 2026
DA - 2026/09/29
PB - arXiv
DO - 10.48550/arxiv.2606.20223
UR - https://arxiv.org/abs/2606.20223
AB - Camera-trap monitoring in African tropical forests increasingly extends beyond closed-canopy interiors to riverbanks, clearings, and park edges. Among available open tools for African forest camera-trap classification, DeepForestVision is the only one providing a matched offline workflow for both photographs and videos, and previous work showed that it outperformed other available baselines on a comparable benchmark. However, it was designed for closed-canopy, ground-level forest interiors and uses a 35-class prediction space that becomes too coarse when deployments encounter arboreal primates, birds, semi-aquatic taxa, or human-associated confounders such as livestock. We present DeepForestVisionV2, an ecology-driven expansion from 35 to 64 prediction classes (61 animal classes plus human, vehicle, and blank) designed to address three recurrent deployment gradients: vertical stratification, scene openness, and anthropogenic interfaces. DeepForestVisionV2 retains the same offline workflow and is trained on 1,535,010 photographs and 243,354 videos from multi-country African tropical-forest projects. Evaluation combines a cross-country cropped-photo validation set, used to assess robustness across sites and camera-trap settings, with three held-out Uganda video benchmarks spanning the targeted gradients. On the validation set, DeepForestVisionV2 reaches 0.86 accuracy, 0.82 macro-F1, and 0.81 balanced accuracy. On the deployment benchmarks, it preserves or improves baseline accuracy despite its harder classification task, while increasing the number of identified taxa from 22 to 29 in forest-interior videos and from 4 to 9 at riverbanks. In the park-edge use case, it raises accuracy from 0.62 to 0.86 and reduces false alarms from 11 to 0. These results show that DeepForestVisionV2 materially improves field utility while preserving robustness across sites, habitats, and camera-trap settings.
ER -