To Do: noise spectrum density estimation and Wiener filtering


Data

To do

  1. For each noise

    • Estimate the spectrum density by periodogram and Welsh method
    • Identify the color of the noise (white, pink, red…)
  2. 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
  3. 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.