Stable-dreamfusion-gui: a GUI for stable-dreamfusion

Content consumption in the 3D space via Extended Reality Technologies is going to be the future, but 3D content creation remains inaccessible to the masses due to its steep learning curve. With recent developments in the latent diffusion text-to-image models field however, researchers have integrate...

Full description

Saved in:
Bibliographic Details
Main Author: Oh, Zhi Hua
Other Authors: Zheng Jianmin
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2023
Subjects:
Online Access:https://hdl.handle.net/10356/171913
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Content consumption in the 3D space via Extended Reality Technologies is going to be the future, but 3D content creation remains inaccessible to the masses due to its steep learning curve. With recent developments in the latent diffusion text-to-image models field however, researchers have integrated well trained image generators for use in 3D objects generation, making 3D content creation far more accessible to content creators. The current best implementation is a Linux-based Command Line Interface application named Stable-DreamFusion, which has limited controls over its Image-to-3D objects generation process. Therefore, this project aims to build a graphical user interface for users to easily convert 2D images to 3D models using the Stable-DreamFusion application, as well as to come up with an alternative workflow to help users generate 3D models that better represent the objects in the input images.