Data processing method and device

A data processing method is applied to the field of artificial intelligence and comprises the steps of obtaining a plurality of first images and random noise; the different first images are denoised images predicted at different historical step lengths through a denoising module in the diffusion mod...

Full description

Saved in:
Bibliographic Details
Main Authors ZHANG SHIFENG, XUE SHUCHEN, YI MINGYANG, LUO WEIJIAN, SUN JIACHENG
Format Patent
LanguageChinese
English
Published 05.12.2023
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:A data processing method is applied to the field of artificial intelligence and comprises the steps of obtaining a plurality of first images and random noise; the different first images are denoised images predicted at different historical step lengths through a denoising module in the diffusion model; and fusing the plurality of first images and the random noise to obtain a denoised image of the current step length. When the denoised image is predicted, the denoised images predicted at different historical step lengths are fused, and the sampling error can be reduced under the condition of the same calculation consumption, so that the sampling frequency is reduced. 一种数据处理方法,应用于人工智能领域,包括:获取多个第一图像和随机噪声;不同的所述第一图像为通过所述扩散模型中的去噪模块在不同历史步长所预测的去噪后的图像;将所述多个第一图像和所述随机噪声进行融合,得到当前步长的去噪后的图像。本申请在预测去噪图像时,将在不同历史步长所预测的去噪后的图像进行融合,可以在同样计算消耗的情况下减小采样误差,从而降低采样次数。
Bibliography:Application Number: CN202310950031