Real-Time Osteoporosis Diagnosis from Knee X-Rays Using anEnhanced YOLO Architecture on Jetson Nano

Abstract Automated osteoporosis detection from knee X-rays poses several challenges, includingsubtle fracture patterns, poor feature representation across scales, and misclassification dueto overlapping bone structures. To overcome these limitations, we propose a novel YOLObased architecture incorporating three key modules: (i) the C3K2 module, which efficientlydownsamples while maintaining robust gradient flow through dual-path processing, (ii) theA2C2F module, which enhances feature learning by splitting inputs into parallel convolution branches before fusion, and (iii) channel-wise concatenation, which merges multi-scalefeatures to improve diversity and spatial coherence without information loss. Experimental evaluation shows that Our model achieves state-of-the-art performance, attaining anmAP@50 of 0.991 for three classes dataset, and mAP50 of 0.892 for two classes dataset,outperforming benchmarks such as YOLO versions. Furthermore, we deployed the modelon an NVIDIA Jetson Nano, demonstrating efficient real-time inference with accurate classification of Normal and Osteoporosis cases. This combination of architectural innovationsand hardware deployment highlights the model’s potential as a portable, reliable solutionfor osteoporosis diagnosis, particularly in low-resource clinical environments.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-17
DOI
https://doi.org/10.21203/rs.3.rs-8446858/v1
Primary Topic
Medical Imaging and Analysis
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Real-Time Osteoporosis Diagnosis from Knee X-Rays Using anEnhanced YOLO Architecture on Jetson Nano

Siddique Ibrahim S P, SUDHIR SANGULA
3 links
Research Square
Medical Imaging and Analysis
preprint

Real-Time Osteoporosis Diagnosis from Knee X-Rays Using anEnhanced YOLO Architecture on Jetson Nano

Siddique Ibrahim S P, SUDHIR SANGULA
preprint en

Abstract

Abstract Automated osteoporosis detection from knee X-rays poses several challenges, includingsubtle fracture patterns, poor feature representation across scales, and misclassification dueto overlapping bone structures. To overcome these limitations, we propose a novel YOLObased architecture incorporating three key modules: (i) the C3K2 module, which efficientlydownsamples while maintaining robust gradient flow through dual-path processing, (ii) theA2C2F module, which enhances feature learning by splitting inputs into parallel convolution branches before fusion, and (iii) channel-wise concatenation, which merges multi-scalefeatures to improve diversity and spatial coherence without information loss. Experimental evaluation shows that Our model achieves state-of-the-art performance, attaining anmAP@50 of 0.991 for three classes dataset, and mAP50 of 0.892 for two classes dataset,outperforming benchmarks such as YOLO versions. Furthermore, we deployed the modelon an NVIDIA Jetson Nano, demonstrating efficient real-time inference with accurate classification of Normal and Osteoporosis cases. This combination of architectural innovationsand hardware deployment highlights the model’s potential as a portable, reliable solutionfor osteoporosis diagnosis, particularly in low-resource clinical environments.

Research Square
VIT-AP University (IN)
Good health and well-being
Medical Imaging and Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon