Tutorial: Self-supervised noise suppression

Published: 01 January 1970
on channel: Software Underground

2,320

Claire Birnie & Sixiu Liu

What you'll need:
Slack channel: #t22-wed-noise-suppression (visit https://softwareunderground.org/slack to join)
Repo: https://github.com/swag-kaust/Transfo...
Pre-requisites: Experience with deep learning and python

Self-supervised learning offers a solution to the common limitation of the lack of noisy-clean pairs of data for training deep learning seismic denoising procedures.

In this tutorial, we will explain the theory behind blind-spot networks and how these can be used in a self-supervised manner, removing any requirement of clean-noisy training data pairs. We will deep dive into how the original methodologies for random noise can be adapted to handle realistic noise in seismic data, both pseudo-random noise and structured noise. Furthermore, each sub-topic presented will be followed by a live, code-along session such that all participants will be able to recreate the work shown and can afterwards apply it to their own use cases.

TIMESTAMPS
00:00:00 Start streaming
00:00:08 Transform 2022 Information
00:01:30 Instructors and Schedule

Presentation. Intro to N2V (Noise to Void)
06:04 Part One - Random noise suppression
07:01 Silence
07:28 Fixed it
11:08 Recap Different Deep learning denoising procedures
11:54 Self-supervised learning examples
13:51 Noise to Void Methodology
17:13 Noise to Void Hyper-parameters

Code
18:49 Tutorial 1 -Random Noise Suppression (N2V with WGN)

Presentation. Adapting to seismic noise
47:30 Part Two - Seismic "random" noise suppression

Code
56:35 Tutorial 2 - Pseudo Random Noise Suppression

Presentation. Coherent noise suppression
1:12:10 Part Three - Coherent noise suppression

Code
1:18:00 Tutorial 3 - Trace-wise Noise Suppression

Presentation
1:36:00 Wrap-up

On this page of the site you can watch the video online Tutorial: Self-supervised noise suppression with a duration of hours minute second in good quality, which was uploaded by the user Software Underground 01 January 1970, share the link with friends and acquaintances, this video has already been watched 2,320 times on youtube and it was liked by 68 viewers. Enjoy your viewing!

9,539

401