Locating the facial region is the aim of face detection in digital images. Face detection issues frequently arise because of digital image noise levels. This study uses median and means filtering techniques to reduce noise in digital photographs. A confusion matrix is used to quantify the median and means filtering methods' accuracy, while the parameters Mean Square Error (MSE) and Peak Noise to Signal Ratio (PNSR) are used to assess these approaches' performance. For this experiment, Viola-Jones was chosen as the face detection method since it is one of the face detection methods with the best accuracy and computational power. According to the outcomes of comparing the median and mean filtering techniques using MSE and PNSR on 50 image samples, the median filtering approach produced the lowest average MSE results, with a value of 19.43, and the median filtering procedure yielded a 13.74 value for the highest PNSR score. The fastest average time was obtained from the mean filtering method with a time of 3.18 seconds. As for the accuracy based on the confusion matrix, these two methods get a good accuracy of 90%. These findings indicate that the Median Filtering approach is superior to the Mean Filtering method in terms of error.