Publications by authors named "Zafran Hussain Shah"
Article Synopsis
- Real-valued convolutional neural networks (RV-CNNs) excel in image restoration tasks like denoising and super-resolution, but they struggle with capturing the full frequency spectrum, leading to potential loss of detail.
- To overcome these limitations, complex-valued convolutional neural networks (CV-CNNs) are explored, which have shown strong performance in other tasks like image classification but are underutilized in image restoration.
- The study introduces new CV-CNN models with complex-valued attention gates for tasks in the frequency domain, demonstrating superior performance in preserving frequency information and achieving better results compared to RV-CNNs.
Background: Convolutional neural network (CNN)-based methods have shown excellent performance in denoising and reconstruction of super-resolved structured illumination microscopy (SR-SIM) data. Therefore, CNN-based architectures have been the focus of existing studies. However, Swin Transformer, an alternative and recently proposed deep learning-based image restoration architecture, has not been fully investigated for denoising SR-SIM images.
View Article and Find Full Text PDF