Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
Thighbone fracture detection based on fused deep learning method
1
Zitationen
2
Autoren
2021
Jahr
Abstract
In this paper, we present an improved deep learning method to detect and localize thighbone fractures in X-ray. Firstly, we fuse feature pyramids with image pyramid to solve the scale variation problem. Secondly, we improve the backbone network in FPN by adding dilated convolutions to stage 5 of ResNet101. To evaluate our method, we establish an annotated dataset including 3842 thighbone fracture X-ray radiographs collected from Linyi People's Hospital. The experiment results show that the trained model achieves the Average Precision (AP) of 87.8% in thighbone fracture detection, and it outperforms other state-of-the-art methods.
Ähnliche Arbeiten
A survey on deep learning in medical image analysis
2017 · 13.953 Zit.
nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation
2020 · 8.117 Zit.
Calculation of average PSNR differences between RD-curves
2001 · 4.093 Zit.
Magnetic Resonance Classification of Lumbar Intervertebral Disc Degeneration
2001 · 3.936 Zit.
Vertebral fracture assessment using a semiquantitative technique
1993 · 3.631 Zit.