Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Skeleton-Guided Diffusion Model for Accurate Foot X-ray Synthesis in Hallux Valgus Diagnosis

Midi Wan, Pengfei Li, Yizhuo Liang, Di Wu, Yushan Pan, Guangzhen Zhu, Hao Wang

TL;DR

This work addresses the need for anatomically accurate foot X-ray synthesis to aid Hallux valgus diagnosis. It introduces the Skeleton-Constrained Conditional Diffusion Model (SCCDM) that operates in a diff-domain and leverages multi-scale feature fusion and self-attention to enforce skeletal integrity while conditioning on natural-light foot images. A new skeletal fidelity metric (KCC) is proposed to quantify keypoint confidence and completeness, and extensive experiments on the Nat2XFoot dataset demonstrate improvements in image quality and anatomical plausibility over strong baselines, with code available at the project GitHub. The approach promises clinically meaningful gains in reliable HV assessment and monitoring by providing high-fidelity, skeletally coherent synthetic X-ray images.

Abstract

Medical image synthesis plays a crucial role in providing anatomically accurate images for diagnosis and treatment. Hallux valgus, which affects approximately 19% of the global population, requires frequent weight-bearing X-rays for assessment, placing additional strain on both patients and healthcare providers. Existing X-ray models often struggle to balance image fidelity, skeletal consistency, and physical constraints, particularly in diffusion-based methods that lack skeletal guidance. We propose the Skeletal-Constrained Conditional Diffusion Model (SCCDM) and introduce KCC, a foot evaluation method utilizing skeletal landmarks. SCCDM incorporates multi-scale feature extraction and attention mechanisms, improving the Structural Similarity Index (SSIM) by 5.72% (0.794) and Peak Signal-to-Noise Ratio (PSNR) by 18.34% (21.40 dB). When combined with KCC, the model achieves an average score of 0.85, demonstrating strong clinical applicability. The code is available at https://github.com/midisec/SCCDM.

Skeleton-Guided Diffusion Model for Accurate Foot X-ray Synthesis in Hallux Valgus Diagnosis

TL;DR

This work addresses the need for anatomically accurate foot X-ray synthesis to aid Hallux valgus diagnosis. It introduces the Skeleton-Constrained Conditional Diffusion Model (SCCDM) that operates in a diff-domain and leverages multi-scale feature fusion and self-attention to enforce skeletal integrity while conditioning on natural-light foot images. A new skeletal fidelity metric (KCC) is proposed to quantify keypoint confidence and completeness, and extensive experiments on the Nat2XFoot dataset demonstrate improvements in image quality and anatomical plausibility over strong baselines, with code available at the project GitHub. The approach promises clinically meaningful gains in reliable HV assessment and monitoring by providing high-fidelity, skeletally coherent synthetic X-ray images.

Abstract

Medical image synthesis plays a crucial role in providing anatomically accurate images for diagnosis and treatment. Hallux valgus, which affects approximately 19% of the global population, requires frequent weight-bearing X-rays for assessment, placing additional strain on both patients and healthcare providers. Existing X-ray models often struggle to balance image fidelity, skeletal consistency, and physical constraints, particularly in diffusion-based methods that lack skeletal guidance. We propose the Skeletal-Constrained Conditional Diffusion Model (SCCDM) and introduce KCC, a foot evaluation method utilizing skeletal landmarks. SCCDM incorporates multi-scale feature extraction and attention mechanisms, improving the Structural Similarity Index (SSIM) by 5.72% (0.794) and Peak Signal-to-Noise Ratio (PSNR) by 18.34% (21.40 dB). When combined with KCC, the model achieves an average score of 0.85, demonstrating strong clinical applicability. The code is available at https://github.com/midisec/SCCDM.
Paper Structure (9 sections, 12 equations, 3 figures, 2 tables, 1 algorithm)

This paper contains 9 sections, 12 equations, 3 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Methods
  3. Denoising Diffusion Generative Models
  4. Skeletal-Constrained.
  5. Comprehensive Evaluation Metrics
  6. Experiments
  7. Nat2XFoot.
  8. Results and Discussion
  9. Conclusion

Figures (3)

  • Figure 1: Overview of the Skeletal-Constrained Conditional Diffusion Model. The diffusion process operates in the diff-domain, defined as the pixel-wise difference between the target X-ray image and the conditional natural light image.
  • Figure 2: U-Net Architecture for Diffusion Models with Multi-Scale and Self-Attention Modules for Noise Prediction.
  • Figure 3: Comparison of X-ray Image Synthesis Methods for Hallux Valgus: Natural Images Translated to X-ray Images.