seismic strong noise attenuation based on diffusion model and principal component analysis（10.1109/TGRS.2024.3355460）. Technically, you should to cite the article if you use this code in your work.

This code is used for noise level assessment and noise attenuation. Since the use of diffusion models for noise processing requires a more accurate grasp of the noise level (to determine the number of diffusion reduction steps), we propose to use principal component analysis for assessment.

run main.m function for validation. You should modify the 'main.m' and './matlab_functions/diffusion_model.m'.

support data formation: xxx.mat with variation name "data" in it (you can use matlab to save it) , and you should name them with numbers, for example, "1.mat, 2.mat……".

validation notes: use matlab generate a list to state t of validation data，which should be named as “t_seq.mat”.

use main.py to training and validation your model, please read notes in main.py

if you need pre-trained model, please contact me ([email protected])