Related Experiment Video
Updated: Jan 9, 2026
06:16
Author Spotlight: Unraveling the Pathogenesis of Age-Related Macular Degeneration and Discovering Potential Therapies
Published on: July 28, 2023
3.0K
S-ResNet-34: small sample-ResNet-34 for predicting cervical degeneration in x-ray image data
Zihan Wei1, Han Wu2, Yifei Xu2
1Department of Bone and Soft Tissue Oncology, Chongqing University Cancer Hospital, Chongqing, 400030, China.
BMC Musculoskeletal Disorders
|December 1, 2025
Summary
A new deep learning model, S-ResNet-34, accurately diagnoses cervical curvature abnormalities from X-ray images. This cost-effective solution improves diagnostic accuracy for conditions like cervical spondylosis.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Deep Learning
Background:
- Cervical physiological curvature abnormalities are common and require accurate diagnostic methods.
- Current diagnostic approaches may lack accuracy or be cost-effective.
- Improved deep learning models can enhance the diagnosis of cervical spondylosis.
Purpose of the Study:
- To develop an improved deep learning model for accurate and cost-effective diagnosis of cervical physiological curvature abnormalities.
- To introduce the S-ResNet-34 model, enhancing nonlinear representation for better diagnostic capabilities.
Main Methods:
- Collected and classified X-ray images from 240 cervical spondylosis patients into normal, straightened, and reversed curvature categories.
- Utilized YOLO-V3 for object detection and data augmentation to create an experimental dataset.
- Developed the S-ResNet-34 model based on ResNet-34 architecture with an integrated learnable weight matrix.
Main Results:
- The S-ResNet-34 model achieved 90.94% accuracy, 85.27% F1 score, and 85.36% recall on the test set.
- Demonstrated superior performance compared to CNN, SVM, RNN, DenseNet, and ResNet-152 models.
- Showcased effectiveness in distinguishing cervical curvature abnormalities in small-sample X-ray image data.
Conclusions:
- The S-ResNet-34 model offers an innovative deep learning approach for auxiliary diagnosis in small-sample X-ray analysis.
- This model simplifies the diagnostic process while maintaining high diagnostic accuracy.
- Presents a new, accurate, and efficient option for diagnosing cervical curvature abnormalities.
