DeweiHu / OCT_DDPM

Application of Diffusion Probablistic Model for unsupervised OCT denoising

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

levels of SNR

chenyzzz opened this issue · comments

Hello! Great job!
I would like to ask you how to use three different levels of signal-to-noise ratio (92dB, 96dB, 101dB) to obtain pictures?
As you said in paper: To test the performance of the model at different speckle levels, the data is acquired with three different levels of signal-to-noise ratio (92dB, 96dB, 101dB).
I want to use my own dataset, but I don't know how to do this step, can you give some advice? Thank you very much!

Thank you for your reply! Please forgive me for having too many questions
Is it possible to add some noise (gaussian noise, salt and pepper noise, speckle noise) to my own dataset and then use your method to denoise? Are there any requirements for the type of these noises? Is it just added in normally?
Thanks again for your reply! !

Hi there, Since we are using our in-house data, the speckle noise level is modulated by attenuating the input signal of the OCT system during acquisition. If you have clean data, you can probably add speckle noise yourself to create noisy dataset. Dewei

________________________________ From: Chenyunzhu @.> Sent: Tuesday, May 30, 2023 10:22 PM To: DeweiHu/OCT_DDPM @.> Cc: Subscribed @.> Subject: [DeweiHu/OCT_DDPM] levels of SNR (Issue #7) Hello! Great job! I would like to ask you how to use three different levels of signal-to-noise ratio (92dB, 96dB, 101dB) to obtain pictures? As you said in paper: To test the performance of the model at different speckle levels, the data is acquired with three different levels of signal-to-noise ratio (92dB, 96dB, 101dB). I want to use my own dataset, but I don't know how to do this step, can you give some advice? Thank you very much! — Reply to this email directly, view it on GitHub<#7>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AL6WZQAWRRFWT5P6XAVEJ7DXI22P3ANCNFSM6AAAAAAYU2V4TE. You are receiving this because you are subscribed to this thread.Message ID: @.>

I get it! I'll try to do it. Thank you.