Publications by authors named "Wangsen Lan"
Article Synopsis
- Residual histograms can offer valuable statistics for low-level visual research, but current image denoising methods often overlook the benefits of using multiple residual histograms for optimization.
- This paper introduces the alternating multiple residual Wasserstein regularization model (AMRW), designed to effectively utilize multiple residual Wasserstein constraints and different image prior knowledge for improved image denoising.
- The AMRW framework enhances noise estimation by aligning the residual histograms of degraded images with a Gaussian noise histogram, allowing for significant improvements in image restoration quality and offering new insights for other visual processing tasks.
