Structural constrained virtual histology staining for human coronary imaging using deep learning
This work addresses the need for real-time histological visualization in coronary artery disease diagnosis, offering a non-invasive alternative, though it appears incremental as it builds on existing GAN methods with structural constraints.
The paper tackled the problem of invasive and time-consuming histology for coronary artery disease by proposing a deep learning method to generate virtual histology staining from OCT images, achieving better performance than conventional GAN-based methods and revealing human coronary layers.
Histopathological analysis is crucial in artery characterization for coronary artery disease (CAD). However, histology requires an invasive and time-consuming process. In this paper, we propose to generate virtual histology staining using Optical Coherence Tomography (OCT) images to enable real-time histological visualization. We develop a deep learning network, namely Coronary-GAN, to transfer coronary OCT images to virtual histology images. With a special consideration on the structural constraints in coronary OCT images, our method achieves better image generation performance than the conventional GAN-based method. The experimental results indicate that Coronary-GAN generates virtual histology images that are similar to real histology images, revealing the human coronary layers.