Student : DA Chan
About the student
I am passionate about applying computational techniques to extract valuable patterns from complex datasets and hope to gain exposure to cutting-edge methodologies in predictive modelling to achieve my objectives.
About the Project
Deraining focuses on developing algorithmic methods to effectively remove rain streaks and artifacts from images while preserving the underlying scene content and image quality. We trace the development from early model-driven approaches that utilized decomposition techniques and dictionary learning, to the emergence of data-driven deep learning methods that now dominate the field. Our comprehensive review categorizes the progression of techniques: from convolutional neural networks, through encoder-decoder architectures, to recent diffusion model-based methods. The results of this study will aim to find the ideal deraining method that makes for more robust object detection.
