Paper
9 January 2025 Application of denoising diffusion probabilistic model in the Minecraft environment
Sheng Li
Author Affiliations +
Proceedings Volume 13486, Fourth International Conference on Computer Vision, Application, and Algorithm (CVAA 2024); 134863A (2025) https://doi.org/10.1117/12.3055770
Event: Fourth International Conference on Computer Vision, Application, and Algorithm (CVAA 2024), 2024, Chengdu, China
Abstract
In this paper, we study the influence of hyperparameters of image generation model in Denoising Diffusion Probabilistic Model (DDPM) on image generation. Experiments were performed on MineRL dataset, the batch size, dimensions, learning rate, and sampling time steps of the model were adjusted, and the Fréchet Inception Distance (FID) was used to evaluate the quality of the resulting images. We introduce a performance degradation index to compare the effects of different hyperparameter settings. The experimental results show that dimension and learning rate are the key factors affecting the quality of DDPM image generation. These findings are of great significance for optimizing the DDPM model and improving its performance in image generation tasks.
(2025) Published by SPIE. Downloading of the abstract is permitted for personal use only.
Sheng Li "Application of denoising diffusion probabilistic model in the Minecraft environment", Proc. SPIE 13486, Fourth International Conference on Computer Vision, Application, and Algorithm (CVAA 2024), 134863A (9 January 2025); https://doi.org/10.1117/12.3055770
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Image quality

Image processing

Diffusion

Denoising

Education and training

Sampling rates

Data modeling

Back to Top