Super-resolution reconstruction of synchrotron radiation nano-CT images and its applications

[Methods]: A Transformer-based U-shaped symmetric network, UTSR, is proposed, integrating a Double Convolutional Feedforward Network (DCFFN) and a Grouped Multi-scale Window Self-Attention (GMSA) mechanism. The network achieves efficient reconstruction through multi-level feature extraction and hierarchical processing. A transfer learning strategy is adopted, leveraging natural image datasets for pre-training and fine-tuning with nano-CT data from the Shanghai Synchrotron Radiation Facility.

A Sobel edge loss function is introduced to prioritize texture preservation, and a two-stage training framework (general image pre-training and nano-CT fine-tuning) optimizes model performance. [Results]: UTSR demonstrates superior performance on both simulated and nano-CT datasets: it achieves lower NIQE (9.38) and LPIPS (0.312) scores compared to SwinIR and Real-ESRGAN, alongside reduced computational complexity (45.72G FLOPs).

When applied to crack segmentation, UTSR-processed images attain an accuracy of 99.3%, with a 24.5% improvement in recall and a 19.1% increase in F1 Score over original images. Ablation experiments confirm the effectiveness of the DCFFN module, where its combination with GMSA further reduces NIQE to 10.99. [Conclusions]: The UTSR network effectively addresses the challenges of low-quality synchrotron radiation nano-CT images and scarce datasets through its innovative architecture and transfer learning strategy, significantly enhancing reconstruction quality and crack segmentation precision.

This research provides a robust preprocessing solution for microstructural analysis and expands the application of deep learning-based super-resolution in scientific imaging, offering substantial academic and engineering value.