SVM-Based Switching Filter Hardware Design for Mixed Noise Reduction in Digital Images Using High-Level Synthesis Tools

Impulse and Gaussian are the two most common types of noise that affect digital images due to imperfections in imaging process, compression, storage communication. The conventional filtering approaches, however, reduce image quality terms sharpness resolution while suppressing effects noise. In this work, a machine learning-based structure has been proposed preserves effectively removing Specifically, support vector classifier is employed detect type affecting each pixel select an appropriate filter. choice filters includes Median Bilateral different kernel sizes. trained using example with known parameters. shown perform better than approaches metrics. Moreover, design implemented as hardware accelerator on FPGA device high-level synthesis tools.