To Do: noise spectrum density estimation and Wiener filtering
Data
To do
-
For each noise
- Estimate the spectrum density by periodogram and Welsh method
- Identify the color of the noise (white, pink, red…)
-
With the image Barbara
- Simulate a noisy version of the image with various SNR (0dB, 5 dB, 10 dB, 15 dB, 20 dB), using a Gaussian white noise
- Denoise the image using the Wiener filter
-
With the noised musics:
- Simulate a noisy version of the music using the three noises using a global factor of $0.1$ for each noise.
- Denoise the sound using the Oracle Wiener filter.