🤖 AI Summary
MRI denoising is hindered by the reliance of supervised learning on paired clean/noisy images, where acquiring ground-truth clean data is prohibitively expensive. To address this, we propose Coil2Coil (C2C), the first fully self-supervised MRI denoising method leveraging multi-channel phased-array coil data. C2C exploits the inherent spatial diversity and statistical independence of noise across coil channels: it constructs pseudo-label pairs via coil-group splitting and recombination—requiring no additional scans or ground-truth images. Integrated with sensitivity normalization and adaptation to the Noise2Noise framework, C2C employs a U-Net architecture for end-to-end training. Evaluated on both synthetic and real DICOM datasets, C2C matches supervised methods in performance while substantially outperforming existing self-supervised approaches. Crucially, its residuals exhibit no structured correlations, confirming effective noise separation. The method is plug-and-play compatible, demonstrating strong potential for clinical deployment.
📝 Abstract
Denoising of magnetic resonance images is beneficial in improving the quality of low signal-to-noise ratio images. Recently, denoising using deep neural networks has demonstrated promising results. Most of these networks, however, utilize supervised learning, which requires large training images of noise-corrupted and clean image pairs. Obtaining training images, particularly clean images, is expensive and time-consuming. Hence, methods such as Noise2Noise (N2N) that require only pairs of noise-corrupted images have been developed to reduce the burden of obtaining training datasets. In this study, we propose a new self-supervised denoising method, Coil2Coil (C2C), that does not require the acquisition of clean images or paired noise-corrupted images for training. Instead, the method utilizes multichannel data from phased-array coils to generate training images. First, it divides and combines multichannel coil images into two images, one for input and the other for label. Then, they are processed to impose noise independence and sensitivity normalization such that they can be used for the training images of N2N. For inference, the method inputs a coil-combined image (e.g., DICOM image), enabling a wide application of the method. When evaluated using synthetic noise-added images, C2C shows the best performance against several self-supervised methods, reporting comparable outcomes to supervised methods. When testing the DICOM images, C2C successfully denoised real noise without showing structure-dependent residuals in the error maps. Because of the significant advantage of not requiring additional scans for clean or paired images, the method can be easily utilized for various clinical applications.
