Enhance Low-Resolution Face Recognition Accuracy Using Median and Mean Filtering Techniques Based on Convolutional Neural Networks (CNN)

The aim of this study is to evaluate the effectiveness of median and mean filtering techniques in enhancing the performance of deep learning-based face recognition systems on low-resolution images. The main challenge in face recognition lies in the low quality of images due to various factors such as lighting, distance, and camera resolution. Median filtering replaces each pixel with the median value of its neighbors, effectively removing salt-and-pepper noise, while mean filtering replaces each pixel with the average value of its neighbors, resulting in smoother images but potentially blurring important details. The results show that median filtering is more effective in preserving edge details and reducing speckle noise, whereas mean filtering excels at producing overall smoother images. The combination of these techniques is expected to enhance the accuracy of face recognition on low-resolution images, contributing to the development of more reliable and precise face recognition systems