Related Experiment Video
Updated: Jun 17, 2026
Automated Joint Space Detection Improves Bone Segmentation Accuracy
Published on: November 28, 2025
Artificial intelligence for detecting subtle paediatric tibial fractures in children under three years: An analysis
1Department of Computer Science, Streatham Campus, University of Exeter, North Park Road, Exeter, EX4 4QF, UK.
Insights
Object detection models like YOLOv8 show promise in identifying subtle tibial fractures in young children from X-rays. Adding a Convolutional Block Attention Module (CBAM) improved accuracy, aiding in early diagnosis and safeguarding.
Area of Science:
- Medical imaging analysis
- Artificial intelligence in radiology
- Pediatric radiology
Background:
- Subtle tibial fractures in children under three are often missed, complicating diagnosis in suspected abuse cases.
- Timely and accurate identification of pediatric tibial fractures is crucial for patient care and protection.
- This study investigates the efficacy of YOLO-based object detection for identifying subtle tibial fractures on pediatric radiographs.
Purpose of the Study:
- To evaluate the capability of YOLO-based object detection models in detecting subtle tibial fractures in pediatric patients.
- To assess the impact of the Convolutional Block Attention Module (CBAM) on model performance.
- To explore the interpretability of the models using Eigen-CAM heatmaps.
Main Methods:
- A dataset of 245 tibial fractures and 637 non-fracture radiographs was retrospectively collected.
- YOLOv5, YOLOv8, and YOLOv11 models were trained and validated using five-fold cross-validation.
- Performance metrics included mAP50, mAP50-95, and per-image diagnostic accuracy, with and without CBAM.
Main Results:
- YOLOv8 demonstrated the highest overall performance in detecting tibial fractures.
- Incorporating CBAM enhanced localization and increased mean detection confidence by 10.6%.
- Buckle fractures were detected most reliably; periosteal and spiral fractures remained challenging.
Conclusions:
- YOLO-based models, particularly YOLOv8 with CBAM, show significant potential for detecting subtle pediatric tibial fractures.
- AI-assisted detection can support radiologists, reduce missed fractures, and improve diagnostic pathways, especially in resource-limited settings.
- Further validation with larger, multicenter datasets is necessary to ensure generalizability.
Introduction:
Subtle tibial fractures in children under three are frequently missed, particularly in cases of suspected physical abuse. Early, accurate identification is essential for clinical decision-making and safeguarding. This study assessed whether YOLO-based object detection models can reliably detect subtle tibial fractures on paediatric radiographs.
Methods:
A retrospective dataset of 245 tibial fractures and 637 non-fracture anterior-posterior radiographs was collected. Fractures were annotated using radiologist-verified polygonal labels. YOLOv5, YOLOv8, and YOLOv11 were trained with five-fold cross-validation, and the effect of adding Convolutional Block Attention Module (CBAM) was examined. Performance was evaluated using mAP50, mAP50-95, and per-image diagnostic accuracy. Eigen-CAM heatmaps were used to assess model interpretability.
Results:
YOLOv8 achieved the strongest overall performance. Adding CBAM improved localisation and increased mean detection confidence by 10.6%. Buckle fractures were most reliably identified, while periosteal and spiral fractures remained challenging, likely due to limited representation and subtle radiographic appearance.
Conclusion:
YOLO-based models show potential for detecting subtle tibial fractures in infants and toddlers, and CBAM further enhances performance in difficult cases. Larger multicentre datasets are needed to strengthen generalisability.
Implications For Practice:
AI-assisted detection could support radiologists in suspected non-accidental injury, reduce missed fractures, and improve diagnostic pathways, particularly in settings without specialist paediatric radiology expertise.