查看文章

The primary motivation behind this article lies in addressing the growing demand for image processing applications within critical environments where the presence of noise can significantly degrade the quality of the outputs. Existing methods primarily consider the noisy version of the image for denoising, neglecting crucial properties such as the dimension of image pixels. This oversight can lead to inaccurate estimates of image values. In the case of Patch-based Local Principal Component Analysis (PL-PCA) filters, the choice of similarity weights and the reduction of sample sizes are pivotal factors. Increasing the sample size can complicate image handling. Regrettably, existing filtering techniques focus solely on noise removal without taking into account the size of input samples. This article aims to resolve these challenges by introducing the H-PL-PCA (Hybrid Patch-based Principal Component Analysis) approach for efficient image noise filtering. The proposed method addresses issues related to dimensionality reduction and the selection of neighboring cells, ensuring that the end results are obtained through the comprehensive analysis of various parameters for effective denoising.