Researching topics for My Thesis (Master's Edition)

Below are some of the ideas that popped up in my mind while thinking about what technology I can work on, which will help me learn and also contribute to the field of Digital Media.

1. Topics:

Streamlining steps to procedurally generate landscapes in Houdini and Unreal Engine by testing tools with the help of AI!

This will help artists work on the parameters to change the landscape, and AI can reduce time by building up the structure and connections required to build that landscape. Whenever I work on any project involving big maps, it makes me feel overwhelmed by how much time it takes to build something from scratch, and procedural generation is one of the solutions that made me feel how much we can improve it just by adding AI. If I could find a tool, the pipeline for developing landscapes would become more streamlined than ever for environmental designers. PCG is research where they use GAN to make landscapes via Height maps. Need to Experiment with AI and work on the concepts to see if the steps will work. Many are working on diverse ways to add AI to PCG, but no one has imagined landscapes.

One of the old projects that made me interested in the topic: https://infinteterrainbytarun.blogspot.com/

Integrating Natural Language Commands with Procedural 3D Generation in Houdini

This thesis aims to create or test ways to implement AI in Houdini, where all types of users can create things without any prior knowledge of the software. Houdini supports Python, which is a language used to create Artificial Intelligence and its database. The technology will help all types of professionals and people who do not know anything about 3D space. I will start targeting people who have tried to test and go to the foundation of how AI works to understand principles to improve it to help creators. I would have to research how NLP works and learn Houdini to get my basis straight on how creators work. The idea is very novel as it's narrowed down to implement it on a scale of software.

Bridge to execute the project: https://github.com/Bismuth-Consultancy-BV/MLOPs

2. Discussion with Other Professionals in the same Media: 

a) According to Professor Emil Polyak, rather than creating an AI for landscape generation, the focus should be on creating a tool with AI that will help creators make VFX for games, animation, or movies. As we are combining PCG, it would be interesting to see what the result would be.

b) After discussing this topic with Professor Michael Wagner, he said that Professor Emil's idea for creating a tool is the best path to go with. He also suggested narrowing the focus down to implementing this idea into Houdini, which will help users who have no prior knowledge and delve deeper into the intricacies of how natural language processing (NLP) can be formed.

c) Professor Debra Issac expressed that this is quite good research, and she will support me in guiding me. She even stated that some of her own students are testing ways to use AI with Houdini, and she will share some resources with me to help me further.

Final Statement and Plans:

Title: Integrating Natural Language Commands with Procedural 3D Generation in Houdini 

Aim   

The aim of this thesis is to explore and develop methods for integrating natural language commands into Houdini's procedural 3D generation system. The objective is to enable users, regardless of their prior experience with 3D software, to create and modify 3D objects by using simple text input. By leveraging Houdini's Python API, this project seeks to demonstrate how natural language processing techniques can bridge the gap between complex 3D modeling tools and user-friendly interaction, making 3D creation more accessible to a wider audience. 

Clarity and Scope  

The focus of the project is both clear and narrow, centering on the application of natural language processing to a specific software—Houdini. The scope is limited to understanding how text-based commands can influence procedural generation and how such a tool can be effectively implemented within Houdini's framework. This project avoids the broad scope of general AI-driven modeling and remains focused on a precise application of language processing to a single software environment. 

Audience 

The primary audience for this research includes professionals in fields like game development, visual effects, and digital art, as well as beginners interested in 3D modeling but lacking the technical skills typically required for tools like Houdini. The research will benefit both experts looking for efficiency improvements and newcomers who want a user-friendly way to engage with 3D creation. This audience is likely to care about the work due to the growing demand for intuitive design tools that lower the technical barrier to entry in 3D content creation.

Research Ability 

This project is researchable through a combination of existing research on natural language processing, procedural content generation, and integration of advanced techniques in digital tools. There is extensive literature on the principles of language processing, such as the use of neural networks and transformer models for understanding text (Vaswani et al., 2017). Additionally, Houdini's existing Python API and procedural workflow offer a concrete platform for experimentation. However, there is limited research specifically on using natural language to directly manipulate 3D spaces in Houdini, which highlights the potential for original contribution. 

Methodology 

The methodology will involve several stages, starting with data collection on commonly used natural language commands and how they correspond to actions within 3D modeling software. Language processing techniques will be utilized to parse text input and map it to relevant Python-based Houdini operations. A prototype tool will be developed, allowing users to input commands and generate or modify 3D objects in real time. Testing will be conducted with both novice and expert users to evaluate the tool's usability, with measurable results based on user satisfaction and task completion time.

Originality 

The originality of this project lies in its attempt to integrate text-based commands into a procedural modeling environment like Houdini, a concept that has not yet been widely explored. While research on language processing and procedural generation exists independently, the combination of these fields in the context of 3D modeling is novel. This project aims to contribute new insights by showing how advanced techniques can make 3D modeling more intuitive, lowering the barrier to entry for non-experts and enhancing workflow efficiency for experienced users.