The Influence of AI on Residential

building design : 3D Rendering

S9 DISSERTATION

ARAVIND K ASHOKAN
ROLL NO : 13
TITLE: The Influence of AI on Residential building design : 3D Rendering

INTRODUCTION:

Understanding AI’s impact is essential for architects to stay relevant, use AI effectively, and address concerns like losing creative control
and ethical issues. AI technologies, such as machine learning, generative design, and advanced visualization tools, have significantly
impacted the architectural landscape by streamlining processes, enhancing creativity, and reducing the workload of architects. By
automating repetitive tasks and providing intelligent insights, AI allows architects to focus on more complex design challenges,
ultimately leading to more innovative and efficient solutions. AI-driven tools can analyze existing structures, assess potential design
modifications, and generate optimized renovation plans that align with the client's vision and budget. This not only accelerates the
planning phase but also enhances the accuracy of project outcomes.

Large Language Models (LLMs) such as OpenAI's GPT-3 and text-to-image models like Stable Diffusion have revolutionized the
potential for generating data. By utilizing ChatGPT and, it is now possible to generate natural-sounding text content and photorealistic
images at an unprecedented scale. These models have proven to be capable of producing high-quality text and images.

Impacts of AI in Architecture Design

• AI generative software can produce a wide range of design solutions based on specific input criteria like material usage, spatial
relationships, budget, sustainability, and more.
• AI can produce high-quality visualizations, 3D models, and even virtual walkthroughs. This helps architects and clients better
understand spatial relationships and design intent.
• AI can assist in projects by quickly analyzing the existing structure, and visualizing possible changes.
HISTORY AND EVOLUTION:
HISTORY AND EVOLUTION:

1. Early Beginnings: Pre-AI (1960s-1980s)

In the early stages, computers were primarily used for basic computations and drafting. Architects used Computer-Aided Design (CAD) software, like
AutoCAD (introduced in the 1980s), which replaced manual drafting with digital drawings.

2. Birth of AI in Architecture: 1990s

The 1990s marked the beginning of more sophisticated computer applications in architecture. Early research focused on automating design tasks using AI
principles, like rule-based systems. For example, programs were developed to help generate basic layouts or optimize spaces based on predefined rules and
constraints

3. Parametric Design and Evolutionary Algorithms: 2000s

In the early 2000s, AI began to play a more significant role with the advent of parametric design. This method allowed architects to create adaptable
designs by defining parameters that could be altered to generate various forms and structures. Programs like Grasshopper for Rhino became popular,
giving architects tools to explore complex geometries and patterns.

4. Machine Learning and Deep Learning: 2010s

AI truly began to mature in the 2010s with the rise of machine learning (ML) and deep learning. These techniques allow computers to "learn" from large
datasets, improving performance without being explicitly programmed. In architecture, this meant AI could analyze vast amounts of design data, identifying
patterns and suggesting design solutions based on past projects.

6. AI in Visualization and Virtual Reality (VR): 2020s

In the 2020s, AI has taken a prominent role in plan generation and visualization. Architects are using AI-powered software for photorealistic rendering and
real-time 3D modeling. AI can predict how light will interact with a space or how materials will look in various environments. Virtual reality (VR) and
augmented reality (AR), often enhanced by AI, allow clients and designers to immerse themselves in virtual spaces and explore designs before they are
built.
STUDY OF AI SOFTWARES
AIM:
The aim is to analyze the influence of Artificial Intelligence (AI) software in the residential building designing process,
specifically focusing on how AI software facilitates and improves the generation of 3d Renders compared to traditional
architectural methods.

OBJECTIVE:
• Background study about the AI software available for 3d Visualization.
• Preparation of questionnaires based on identified parameters.
• Comparative analysis of AI-driven design process and traditional methods by using the parameters.
• Analyze the influence of AI software in the residential building designing process based on the comparative analysis
and arriving at the conclusion.

SCOPE:
• Future possibilities for AI integration in the residential architecture sector.
• The application of AI in residential building design, particularly in the areas of 3D visualization.

LIMITATIONS:
• Access to detailed case studies and data on AI-driven residential projects may be limited, which could affect the
depth of analysis.
• The scope is limited to residential buildings and does not extend to commercial or industrial renovation
processes.
LITERATURE REVIEW :

PARAMETERS IDENTIFIED : SOFTWARE IDENTIFIED :

• WORKING TIME • STABLE DIFFUSION
• FLEXIBILITY
• LIGHTING
• MATERIAL REPRESENTATION
• DESIGN COMPLEXITY
METHODOLOGY:
ABOUT THE AI SOFTWARE :

Stable Diffusion is a cutting-edge AI-driven image generation model that leverages deep learning techniques to create high-quality,
realistic images. Stable Diffusion is designed to generate a wide range of images, from photorealistic photos to abstract art, making it
a versatile tool for artists, designers, and architects. By training on vast datasets of images, the model learns to generate new images
that maintain the intricate details and characteristics of the original data. This results in outputs that are not only visually striking
but also contextually accurate.

Stable Diffusion has two basic features: text-to-image and image-to-image. Its text-to-image feature can transform parameter-
triggered descriptive prompts into high-quality, creative, and unique images. Its image-to-image feature is designed to generate a
new image based on the images you enter as input. One of the key advantages of Stable Diffusion is its flexibility. Users can input
simple text prompts or rough sketches, and the model will generate detailed images that match the given input.

Key features :-

•Most rapid image generator, without compromising the quality of the images
•Open-source, users can access and modify the source code and model the tool as required.
•Can be personally trained by feeding the AI specific datasets to generate images that are aligned with their preferences.
•Advanced technical features, like negative prompting and options to fine-tune the model.
QUESTIONNAIRE FOR SURVEY :
WORKING TIME
• How long does it typically take you to generate a detailed render of a residence?
a) Less than 10 min
b) 10 – 20min
c) 20 - 40 min
d) More than 40 min

• In your opinion which method (AI or human architect) produces a basic concept or draft design faster?
a) AI-driven design
b) Architect
c) Both
d) Neither
• In your opinion which method allows for quicker revisions when changes are needed?
a) AI-driven design
b) Architect
c) Both
d) Neither

FLEXIBILITY
• How often do you make changes to your initial design plan based on client feedback? • In your opinion which design process (AI or architect) provides greater
a) Rarely customization options to meet unique client preferences?
b) Sometimes a) AI-driven design
c) Often b) Architect
d) Always c) Both
d) Neither
• In your opinion which method is more flexible when responding to sudden design changes during
the project?
a) AI-driven design
b) Architect
c) Both
d) Neither
QUESTIONNAIRE :

LIGHTING
• In your opinion which method provides more accurate and aesthetically pleasing lighting effects in the 3D rendering?
a) AI-driven software
b) Human architect
c) Both
d) Not aware
• In your opinion which method better adjusts lighting effects based on different environments (e.g., indoor, outdoor, urban,
rural)?
a) AI-driven design
b) Human architect
c) Both
d) Not aware

DESIGN COMPLEXITY
• In your opinion which method can better handle complex geometrical forms, such as irregular
shapes or advanced structural geometries?
a) AI-driven design
b) Human architect
c) Both
d) Not aware
• Which method (AI or conventional) handles multiple iterations of a highly complex design more
efficiently?
a) AI-driven design
b) Human architect
c) Both
d) Not aware
QUESTIONNAIRE :
MATERIAL REPRESENTATION
• In your opinion which design method optimizes material textures better for both efficiency and
user experience?
a) AI-driven design
b) Human architect
c) Both
d) Not aware

• In your opinion which method captures the surface qualities of materials (e.g., roughness,
gloss, transparency) more realistically?
a) AI-driven design
b) Human architect
c) Both
d) Not aware

• In your opinion which method better represents a wider variety of materials and textures (e.g., wood,
concrete, fabric) in the 3D rendering?
a) AI-driven design
b) Human architect
c) Both
d) Not aware
QUESTIONNAIRE FOR ARCHITECT :
WORKING TIME
• How long does it typically take you to produce a high-quality rendering for a standard project?
• Have you noticed any specific stages of the rendering process that are significantly faster since adopting AI?

FLEXIBILITY
• How often do you make changes to your initial design plan based on client feedback?

• Do you find AI makes it easier or more difficult to introduce creative, last-minute changes?

LIGHTING
• Have you noticed improvements in the way AI handles dynamic lighting scenarios, such as changing between day and night?

• Do you still need to manually adjust lighting for more nuanced atmospheric effects, or does AI handle this well?

DESIGN COMPLEXITY
• How well does Stable Diffusion handle complex architectural designs, such as irregular geometries or multi-functional spaces, compared to traditional methods?
• Do you find that the AI software enhances your ability to manage design complexity, or does it require additional oversight to ensure accuracy?

MATERIAL REPRESENTATION
• Is it easier to experiment with different material finishes in Stable Diffusion?
• Has the accuracy of material reflection and transparency improved with AI?
QUESTIONNAIRE FOR ARCHITECT
• Do you feel that AI has improved your overall design efficiency, or has it introduced new challenges in the workflow?

• Do clients respond differently to AI-generated renderings compared to traditional ones?

• Do you find the AI software to be a complementary tool in your rendering process, or has it replaced most manual tasks?

• Would you ever consider going back to fully manual rendering for specific projects, or has AI completely replaced those methods?

• Have you noticed any significant differences in the final output between AI-generated renderings and those produced manually?

ARCHITECT’S CONCEPTION

Residence 1

Conventional software AI software (Stable Diffusion)

 Conventional software AI software (Stable Diffusion)

Residence 2

Residence 3
Residence 4

Conventional software AI software (Stable Diffusion)

ARCHITECT’S PERCEPTION ANALYSIS

In the Architect’s perspective,

• The use of AI in residential building renderings have impacted their total work time by a satisfactory level.
• The AI software uses easier ways to render the last minute changes according to the client’s feedback.
• The Lighting settings in conventional software can handle lights more precisely by adjusting the position and angle of light and shadow.
• Both software handles complex architectural designs, such as irregular geometric shapes, and exterior detailing, but AI can visualize it quicker.
• Material quality and textures should be manually done with conventional software for better quality settings.
PRACTICAL ANALYSIS:
A practical comparison was done using Stable Diffusion, by training Primary and Secondary model with proper prompt and negative prompt, to compare it
with the render using a Conventional software.

PROMPT :- "A serene evening scene in a modern suburban neighborhood. The sky glows with soft hues of purple, orange, and pink as the sun sets behind
distant hills. Street lamps cast a warm, golden light on the quiet street, lined with sleek contemporary houses. Shadows stretch across neatly manicured
lawns, and a few trees gently sway in the light breeze. The windows of the houses glow with warm interior lighting, hinting at the cozy atmosphere inside.
In the foreground, a modern car is parked on the driveway, reflecting the evening light.“

NEGATIVE PROMPT :- "blurry, out of focus, abstract, unrealistic lighting, overexposed, cartoonish, over-saturated colors, grainy textures, low resolution,
excessive fog, unnatural shadows, glowing objects, exaggerated features, fantasy elements, sci-fi, people or characters, unrealistic buildings, distorted
reflections."

These prompts was given to the image to image generator in Stable Diffusion,
attached with a Base Render of the 3D model. This base render is used for
comparing the rendering potential of the conventional software and AI based
software.

Base Render
PRACTICAL ANALYSIS:

Rendered using a Conventional rendering software Rendered using a AI driven software (Stable Diffusion)

ANALYSIS

After doing a comparative analysis,

• AI in residential building renderings have drastically reduced the time to take a render of same quality as a conventional software.
• The AI software can easily render more options at the lesser time than conventional software.
• The Lighting in conventional software can handle lights more precisely by adjusting the position and angle of light and shadow.
• The AI software visualizes complex geometric shapes and other detailing quicker than conventional software.
• Conventional software offers more precise material and texture customisation through individual material edit options.
SURVEY ANALYSIS
WORKING TIME

ANALYSIS
• AI software is seen as more efficient, with most respondents using AI able to
generate renders within 10-20 minutes.

• AI-driven design is seen as faster for producing basic concepts or draft

designs
SURVEY ANALYSIS
FLEXIBILITY

ANALYSIS
• AI software is widely seen as offering quicker revisions when changes
are needed.

• AI-driven tools were seen as more client responsive due to their

flexibility in handling iterative changes efficiently.

• Conventional software was seen as less flexible, requiring more

manual input and time to implement changes.
SURVEY ANALYSIS
LIGHTING ANALYSIS
• AI software is able to simulate complex lighting conditions, such as
daylight, shadows, and reflections, with higher precision.

• AI tools were generally preferred for their efficiency in automatically

adapting to environmental conditions.
SURVEY ANALYSIS
MATERIAL REPRESENTATION

ANALYSIS
• In terms of representing a wider variety of materials and textures (e.g.,
wood, concrete, fabric), AI was generally favored for its speed and
accuracy.

• AI software received high marks for its ability to represent and optimize
material textures efficiently.
SURVEY ANALYSIS
DESIGN COMPLEXITY ANALYSIS
• AI software was rated more favorably for handling complex geometries,
such as irregular shapes or advanced structural forms, with less manual
input.

• Conventional software is still favored by some respondents for offering

greater control and precision, especially in high-end, bespoke designs
where manual adjustments are important.

• AI software was overwhelmingly rated as more efficient in handling

multiple iterations of complex designs.
RESULT :
From the three analysis done,

• AI-based rendering solutions have shown a significant reduction in total work time for creating renders of comparable quality to conventional
methods.

• The accelerated processing capabilities of AI allow architects to produce multiple rendering options in a shorter timeframe, thus streamlining the
design decision-making process.

• The use of AI allows for more flexibility when responding to client feedback, particularly for last-minute changes. The AI tools can adapt and re-render
updated versions more quickly, minimizing delays. Conventional software may require a full re-rendering cycle, which adds to the total project time.

• Traditional software allows for more precise control over lighting, enabling architects to fine-tune light positioning, shadow details, and angle
adjustments. This higher degree of customization makes it preferable in scenarios where detailed lighting effects are crucial.

• While traditional software is capable of handling complex designs, it requires more processing time, especially when dealing with irregular geometry
or advanced exterior detailing. The visual quality may be superior but is balanced by the longer time investment. The AI software can visualize and
generate these complex designs almost instantaneously, enhancing productivity.

• AI-based software automates much of the rendering process but, material quality and texture settings are generally less customizable and precise
than in conventional software. Traditional tools offer more customization options for materials and textures, enabling architects to fine tune individual
surface details to achieve the highest quality render.
THANK YOU…