A deep learning framework for bone fragment classification in owl pellets using YOLOv12

Abstract Non-invasive monitoring of small mammal populations is critical for both biodiversity conservation and integrated pest management, particularly in agroecosystems. Barn owl (Tyto alba) pellet analysis has long served as a valuable tool for inferring prey abundance, yet conventional bone clas...

Full description

Saved in:
Bibliographic Details
Main Authors: Nik Fadzly, Lay Wai Kean, Siti Nuramaliati Prijono, Rini Rachmatika, Siti Zulaika, Mohd Nasir, Hasber Salim
Format: Article
Language:English
Published: Nature Portfolio 2025-08-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-15906-9
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849332998075318272
author Nik Fadzly
Lay Wai Kean
Siti Nuramaliati Prijono
Rini Rachmatika
Siti Zulaika
Mohd Nasir
Hasber Salim
author_facet Nik Fadzly
Lay Wai Kean
Siti Nuramaliati Prijono
Rini Rachmatika
Siti Zulaika
Mohd Nasir
Hasber Salim
author_sort Nik Fadzly
collection DOAJ
description Abstract Non-invasive monitoring of small mammal populations is critical for both biodiversity conservation and integrated pest management, particularly in agroecosystems. Barn owl (Tyto alba) pellet analysis has long served as a valuable tool for inferring prey abundance, yet conventional bone classification is labour-intensive and requires specialized expertise. Here, we introduce a deep learning framework that automates the detection and classification of rodent bone fragments from owl pellets using the YOLOv12 object detection architecture. A dataset comprising 978 annotated images, encompassing skull, femur, mandible, and pubis bones, was used to train and validate the model, achieving high detection performance (precision = 0.90, recall = 0.90, mAP@0.5 = 0.984, F1-score = 0.97). The model demonstrated strong generalization across samples from Malaysia and Indonesia. We further developed a Python-based inference script to estimate rodent abundance using skull and paired bone counts. This AI-assisted workflow reduces human error, increases processing throughput, and enables scalable rodent monitoring. By enhancing ecological inference from pellet studies, our approach supports timely biodiversity assessments and pest surveillance strategies across diverse landscapes.
format Article
id doaj-art-daae98eeb8f74d8b9ad38b86b0adcf47
institution Kabale University
issn 2045-2322
language English
publishDate 2025-08-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-daae98eeb8f74d8b9ad38b86b0adcf472025-08-20T03:46:01ZengNature PortfolioScientific Reports2045-23222025-08-011511910.1038/s41598-025-15906-9A deep learning framework for bone fragment classification in owl pellets using YOLOv12Nik Fadzly0Lay Wai Kean1Siti Nuramaliati Prijono2Rini Rachmatika3Siti Zulaika4Mohd Nasir5Hasber Salim6School of Biological Sciences, Universiti Sains Malaysia, USMSchool of Biological Sciences, Universiti Sains Malaysia, USMApplied Zoology Research Center, National Research and Innovation Agency (BRIN)Applied Zoology Research Center, National Research and Innovation Agency (BRIN)School of Biological Sciences, Universiti Sains Malaysia, USMSchool of Biological Sciences, Universiti Sains Malaysia, USMSchool of Biological Sciences, Universiti Sains Malaysia, USMAbstract Non-invasive monitoring of small mammal populations is critical for both biodiversity conservation and integrated pest management, particularly in agroecosystems. Barn owl (Tyto alba) pellet analysis has long served as a valuable tool for inferring prey abundance, yet conventional bone classification is labour-intensive and requires specialized expertise. Here, we introduce a deep learning framework that automates the detection and classification of rodent bone fragments from owl pellets using the YOLOv12 object detection architecture. A dataset comprising 978 annotated images, encompassing skull, femur, mandible, and pubis bones, was used to train and validate the model, achieving high detection performance (precision = 0.90, recall = 0.90, mAP@0.5 = 0.984, F1-score = 0.97). The model demonstrated strong generalization across samples from Malaysia and Indonesia. We further developed a Python-based inference script to estimate rodent abundance using skull and paired bone counts. This AI-assisted workflow reduces human error, increases processing throughput, and enables scalable rodent monitoring. By enhancing ecological inference from pellet studies, our approach supports timely biodiversity assessments and pest surveillance strategies across diverse landscapes.https://doi.org/10.1038/s41598-025-15906-9Artificial intelligenceOwl pelletsBone classificationMachine learningObject detection
spellingShingle Nik Fadzly
Lay Wai Kean
Siti Nuramaliati Prijono
Rini Rachmatika
Siti Zulaika
Mohd Nasir
Hasber Salim
A deep learning framework for bone fragment classification in owl pellets using YOLOv12
Scientific Reports
Artificial intelligence
Owl pellets
Bone classification
Machine learning
Object detection
title A deep learning framework for bone fragment classification in owl pellets using YOLOv12
title_full A deep learning framework for bone fragment classification in owl pellets using YOLOv12
title_fullStr A deep learning framework for bone fragment classification in owl pellets using YOLOv12
title_full_unstemmed A deep learning framework for bone fragment classification in owl pellets using YOLOv12
title_short A deep learning framework for bone fragment classification in owl pellets using YOLOv12
title_sort deep learning framework for bone fragment classification in owl pellets using yolov12
topic Artificial intelligence
Owl pellets
Bone classification
Machine learning
Object detection
url https://doi.org/10.1038/s41598-025-15906-9
work_keys_str_mv AT nikfadzly adeeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT laywaikean adeeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT sitinuramaliatiprijono adeeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT rinirachmatika adeeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT sitizulaika adeeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT mohdnasir adeeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT hasbersalim adeeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT nikfadzly deeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT laywaikean deeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT sitinuramaliatiprijono deeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT rinirachmatika deeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT sitizulaika deeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT mohdnasir deeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12
AT hasbersalim deeplearningframeworkforbonefragmentclassificationinowlpelletsusingyolov12