Gaussian Noise Removal in Handloom Images via Edge-Adaptive Total Variation Model

In digital images eradicating noise plays a significant role in image processing and image analysis. Even though a wide range of techniques proposed for the intent of noise elimination in digital images, the problem still remains an open challenge in research. For a framework like an automatic handloom system where a deep learning-based network for design generation and defect detection needs a standard database for training. To produce quality output, the training input images need to be pre-processed. This paper implemented a strong edge preservation, and variational attribute model for reducing noise accordant with an edge-adaptive total variational model. The findings of the study indicate that the approach not only completely eliminates Gaussian noise, but adequately maintains the principal edge content. The performances of the method are assessed on the basis of metrics like Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM) evaluated in contrast to other existing methods.