DESIGN DREAM

AN AI-POWERED HOUSE DESIGN APP

Visualisation and Customization
A Real Time research project
Submitted to
JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY, HYDERABAD
In partial fulfilment of the requirements for the award of degree of

BACHELOR OF TECHNOLOGY
In
COMPUTER SCIENCE AND ENGINEERING
DATA SCIENCE
Submitted by
BOMMERA KEERTHANA (22UK1A6713)
Under the guidance of
Dr. K. SHARMILA REDDY
HOD-CSE (Data science)

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

(DATA SCIENCE)
VAAGDEVI ENGINEERING COLLEGE
BOLLIKUNTA, WARANGAL (T.S)-506005. Affiliated to JNTUH, HYDERABAD
 DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
DATA SCIENCE
VAAGDEVI ENGINEERING COLLEGE
BOLLIKUNTA, WARANGAL-506005

CERTIFICATE

This is to certify that the Real time research project entitled Design Dream an
AI-Powered house design application is being submitted by BOMMERA
KEERTHANA (22UK1A6713) for the partial fulfilment of the requirements
for the award of the degree of Bachelor of Technology in Computer Science and
Engineering to Jawaharlal Nehru Technological University Hyderabad during
the academic year 2023-2024.

Internal Faculty Project Guide HOD

(NAME) Dr. K. Sharmila Reddy Dr. K. Sharmila Reddy
 ACKNOWLEDGEMENT

I wish to take this opportunity to express my sincere gratitude and deep sense of respect to
Dr. Syed Musthak Ahmed, Principal, Vaagdevi Engineering College for making us available
all the required assistance and his support and inspiration to carry out this Real Time research
project at the institute.

I extend my heartfelt thanks to my guide Dr. K. Sharmila Reddy, Head, Department of CSE
(Data Science), Vaagdevi Engineering College for providing me necessary infrastructure and
thereby giving me freedom to carry out technical seminar.

Finally, I express my sincere thanks and gratitude to my family members , friends for their
encouragement and outpouring their knowledge and experiencing throughout thesis.

BOMMERA KEERTHANA (22UK1A6713)

 ABSTRACT
The AI-Powered House Design Application represents a groundbreaking advancement in the
field of architectural design and home customization. Traditional approaches to home design
often involve complex workflows, manual planning, and limited options for personalization.
In contrast, our application harnesses the latest advancements in artificial intelligence (AI)
and real-time visualization technologies to provide users with a seamless and intuitive
platform for creating personalized house layouts.

At its core, the AI-Powered House Design Application combines sophisticated AI algorithms
with interactive 3D visualization tools to offer users an unprecedented level of flexibility,
creativity, and control over the design process. By leveraging machine learning techniques,
the application analysis user preferences, lifestyle requirements, and architectural constraints
to generate intelligent design recommendations tailored to each individual user.

Through a user-friendly interface, users can easily customize and visualize their dream home
in real-time, exploring different layout configurations, room dimensions, and interior design
elements. The application facilitates collaboration and communication between homeowners,
architects, and interior designers, streamlining the design workflow and enabling seamless
integration of user feedback throughout the process.
 TABLE OF CONTENTS

1. INTRODUCTION

2. EXISTING SYSTEM

3. PROPOSED SYSTEM

4. PROBLEM STATEMENT

5. DATA COLLECTION & ANALYSIS

6. DESIGN AND MODELLING

7. MODEL BUILDING / ARCHITECTURE

8. CONCLUSION

9. REFERENCES
 INTRODUCTION
In the contemporary world, home design has evolved beyond basic functionality to
encompass aesthetic appeal, personal expression, and optimized spatial utilization. However,
the traditional home design process can be daunting, often requiring substantial time,
expertise, and financial resources. Many individuals find it challenging to visualize their
ideas and turn them into reality without professional assistance. This gap between concept
and realization highlights the need for innovative solutions that make home design accessible,
intuitive, and efficient.

"DesignDream" is an AI-powered house design application developed to address these

challenges. By integrating advanced machine learning algorithms with a user-friendly
interface, Design Dream enables users to effortlessly create, visualize, and customize their
home designs. The application caters to a broad audience, including homeowners, interior
designers, and real estate professionals, offering a versatile tool that can accommodate
diverse design needs and preferences.

Objectives of Design Dream

1. Simplify the Design Process: Automate complex design tasks to make the process
more manageable for users without professional expertise.
2. Enhance Creativity: Provide users with a platform that encourages creativity by
offering various design options and styles.
3. Personalization: Generate customized design recommendations based on user inputs,
such as room dimensions, type, and preferred styles.
4. Accessibility: Make high-quality design tools available to a wider audience, reducing
the need for professional intervention.
5. Efficiency: Streamline the design workflow to save time and effort for users.
Key Features

 User Registration and Authentication: Secure user accounts with registration and
login functionalities.
 Interactive Design Form: A user-friendly interface to input design preferences and
requirements.
 AI-Powered Design Generation: Advanced algorithms to generate personalized
design suggestions.
 Visualization Tools: High-quality renderings and visualizations to help users see
their designs come to life.
 Customization Options: Flexible tools for users to tweak and refine generated
designs according to their preferences.

Significance

Design Dream represents a significant advancement in the home design industry by bridging
the gap between professional design services and DIY enthusiasts. It democratizes access to
sophisticated design tools, allowing users to experiment with different styles, layouts, and
features without needing specialized knowledge or skills. The application not only saves time
and reduces costs but also enhances user satisfaction by turning their unique visions into
reality.
 EXISTING SYSTEM
1. Planner 5D
 Overview: Planner 5D is an online platform that enables users to create
detailed home and interior designs in 2D and 3D. It offers a user-friendly
interface and a vast library of design elements.

2. Roomstyler 3D Home Planner

 Overview: Roomstyler is an easy-to-use online tool for interior design. It
allows users to create room designs and experiment with furniture and
decoration in a 3D space.

3. Homestyler
 Overview: Homestyler is an interior design application that provides users
with tools to create detailed home designs and visualizations.

4. AutoDesk Revit
 Overview: Revit is a professional-grade Building Information Modeling
(BIM) software used by architects, engineers, and construction professionals
for designing buildings and structures.

5. Houzz
 Overview: Houzz is a platform that connects homeowners with interior
designers and architects. It provides tools for home design and remodeling.
 PROPOSED SYSTEM
1. Customization Level:

 Existing Projects: Varies from basic to very high (Revit), with some offering
extensive libraries and customization options (Planner 5D, Homestyler).
 Design Dream: Provides very high customization levels, leveraging advanced AI
to allow users to customize every aspect of their design in detail.

2. AI Integration:

 Existing Projects: Mostly limited to recommendations and layout suggestions.

Some (like Revit) use AI for performance analysis.
 Design Dream: Uses advanced AI for detailed customization, real-time
visualization, and personalized recommendations, offering a seamless and
intelligent design experience.

3. User Interface:

 Existing Projects: Range from user-friendly (Planner 5D, Roomstyler) to

complex (Revit).
 Design Dream: Focuses on a user-friendly and intuitive interface that makes
complex design tasks simple and accessible.

4. Design Library:

 Existing Projects: Extensive libraries with varying degrees of customizability.

 Design Dream: Offers an extensive and customizable library, enhanced by AI-
driven recommendations for design elements.
5. Visualization Quality:

 Existing Projects: High-quality visualizations but vary in real-time rendering

capabilities.
 Design Dream: Emphasizes high-quality, real-time rendering to provide instant
feedback on design changes.

6. Recommendations:

 Existing Projects: Basic to personalized AI-based recommendations.

 Design Dream: Offers highly personalized and detailed AI-driven
recommendations, adapting to user preferences and design trends.

7. Collaboration Features:

 Existing Projects: Limited in most tools, extensive in professional tools like

Revit.
 Design Dream: Provides moderate collaboration features, allowing users to share
and collaborate on designs easily.
 PROBLEM STATEMENT
Homeowners and designers often face challenges when it comes to envisioning and
customizing home designs. Traditional design processes are time-consuming, require
professional expertise, and often lack interactive visualization tools. Additionally, users may
struggle to find design recommendations that align with their personal tastes and
requirements. Traditional approaches to house design and customization often involve
cumbersome workflows, limited personalization options, and a lack of efficient
communication between stakeholders. Homeowners face challenges in translating their vision
into reality, while architects and interior designers grapple with time-consuming manual
processes and difficulty in understanding and meeting client preferences. This results in
inefficiencies, delays, and dissatisfaction with the final design outcome.

Key Challenges

1. Time-Consuming Design Process: Creating a home design from scratch can be

labour-intensive and requires a significant investment of time.
2. Lack of Visualization Tools: Many users find it difficult to visualize their ideas in a
concrete form, making it hard to make decisions about layouts, colour schemes, and
furnishings.
3. Need for Professional Expertise: The traditional design process often necessitates
hiring professional designers, which can be costly.
4. Personalization and Recommendations: Users need a solution that not only allows
customization but also provides intelligent recommendations that suit their
preferences and needs.
 DATA COLLECTION & ANALYSIS
Data collection and preparation are critical steps in developing the "DesignDream" AI-
powered home design application. The quality of the data significantly impacts the
performance of the AI models and the overall user experience.

Data Collection

1. User Data Collection:

 Demographic Information: Collect information such as age, gender,
location, and occupation to understand the user base.
 Design Preferences: Gather data on user preferences, including favourite
styles, colour schemes, room layouts, and preferred furniture types.
 Interaction Data: Track user interactions within the app to understand how
users navigate, what features they use the most, and their design choices.
2. Design Data Collection:
 Architectural Plans: Collect a variety of architectural plans and blueprints to
understand different room layouts and structures.
 Furniture and Decor: Compile a dataset of furniture and decor items,
including images, descriptions, dimensions, and styles.
 Material and Colour Palettes: Gather data on different materials (wood,
metal, fabric) and colour palettes used in interior design.
3. Market Trends Data:
 Trend Analysis: Use web scraping techniques to collect data from design
magazines, blogs, and social media platforms to identify current trends in
home design.
 Sales Data: Collect sales data from furniture and decor retailers to understand
popular items and styles.
4. External Data Sources:
 Open Data Sets: Utilize open data sets from design-related databases and
repositories.
 Partnerships: Collaborate with design firms and retailers to access
their data.
Data Preparation

1. Data Cleaning:
 Remove Duplicates: Ensure there are no duplicate entries in the data set.
 Handle Missing Values: Address missing data by filling in gaps with
reasonable estimates or removing incomplete records.
 Correct Errors: Identify and correct any errors in the data (e.g., incorrect
dimensions, mislabelled images).
2. Data Transformation:
 Normalization: Normalize the data to ensure consistency, such as
standardizing units of measurement.
 Categorization: Categorize design elements into groups (e.g., modern,
traditional, minimalist) to facilitate easier recommendations and
customizations.
 Encoding: Convert categorical data into numerical values for use in machine
learning models.
3. Data Annotation:
 Labelling: Label images and design elements with relevant tags and metadata
to improve searchability and AI training.
 Feature Extraction: Extract key features from the data, such as room
dimensions, furniture styles, and material types, for use in AI models.
4. Data Splitting:
 Training and Test Sets: Split the data into training and test sets to evaluate
the performance of the AI models. Typically, 70-80% of the data is used for
training, and 20-30% is reserved for testing.
5. Data Augmentation:
 Synthetic Data Generation: Create synthetic data to augment the training set,
such as generating variations of room layouts or furniture arrangements using
existing data.
6. Integration:
 Database Setup: Organize the data into a structured database, such as a SQL
or NoSQL database, to facilitate efficient access and retrieval by the
application.
Data Analysis

Data analysis plays a crucial role in understanding patterns, trends, and insights within the
collected data for an AI-powered house design application. Here's how data analysis can be
conducted:

1. Exploratory Data Analysis (EDA):

 Perform EDA to gain insights into the characteristics and distribution of the
data.
 Explore descriptive statistics, such as mean, median, standard deviation, and
percentiles, to understand the central tendency and variability of numerical
features.
 Visualize data distributions using histograms, box plots, and density plots to
identify outliers, skewness, and other anomalies.
 Analyse categorical features through frequency tables, bar charts, and pie
charts to understand the distribution of different categories.
2. Correlation Analysis:
 Explore the relationships between different features using correlation analysis.
 Calculate correlation coefficients (e.g., Pearson correlation) to quantify the
strength and direction of linear relationships between numerical features.
 Visualize correlation matrices using heatmaps to identify patterns and
dependencies among variables.
3. Feature Importance Analysis:
 Determine the importance of features in predicting design preferences or
outcomes.
 Use techniques such as feature importance scores from machine learning
models (e.g., decision trees, random forests) or permutation importance to
assess the impact of individual features on the target variable.
 Visualize feature importance rankings using bar charts or feature importance
plots to prioritize influential features for model development.
4. Cluster Analysis:
 Conduct cluster analysis to identify groups or clusters within the data based on
similarities or patterns.
  Utilize clustering algorithms (e.g., K-means clustering, hierarchical clustering)
to partition the data into distinct groups.
 Visualize clusters using scatter plots or dendrograms to understand the
underlying structure of the data and potential segmentation patterns.

FEATURES
Here are some key features that can be considered in such an application:

1. Room Dimensions:
 Length, width, and height of each room in the house layout.
2. Room Types:
 Classification of rooms based on their function (e.g., living room, bedroom,
kitchen, bathroom).
3. Furniture Placement:
 Coordinates or positions of furniture items within each room.
4. Interior Design Styles:
 Categorization of design styles (e.g., modern, traditional, minimalist) based on
colour schemes, furniture types, and decor elements.
5. User Preferences:
 Individual preferences of users, including design styles, colour preferences,
furniture preferences, and lifestyle requirements.
6. Spatial Relationships:
 Proximity and adjacency relationships between rooms and furniture items
within the house layout.
7. Natural Lighting:
 Assessment of natural lighting conditions within each room based on window
placement and orientation.
8. Budget Constraints:
 User-defined budget constraints for furniture purchases, renovation costs, or
overall project expenses.
9. Energy Efficiency:
 Evaluation of energy-efficient design features, such as insulation, HVAC
systems, and renewable energy sources.
10. Environmental Factors:
 Consideration of environmental factors, such as climate conditions,
geographic location, and sustainability principles.

1. Visualization:

 3D Rendering: Utilize advanced 3D rendering techniques to create realistic and

immersive visualizations of house layouts and interior design elements.
 Interactive Controls: Implement interactive controls that allow users to navigate
through the 3D environment, zoom in/out, rotate, and pan to explore different
perspectives.
 Real-Time Updates: Ensure that visualizations are updated in real-time as users
make changes to the house layout or design elements, providing instant feedback and
enhancing the user experience.
 Lighting and Textures: Incorporate realistic lighting effects and textures to simulate
natural lighting conditions and material finishes, enhancing the visual appeal of the
designs.

2. Customization:

 Drag-and-Drop Interface: Provide a user-friendly interface that allows users to drag

and drop rooms, furniture items, and design elements onto the house layout canvas,
enabling intuitive customization.
 Flexible Layout Editing: Allow users to resize, rotate, and reposition rooms and
furniture items within the layout canvas, giving them full control over the design
process.
 Design Templates: Offer pre-designed templates or layouts that users can customize
and adapt to their preferences, providing inspiration and guidance for the design
process.
 Material and Colour Selection: Enable users to choose from a variety of materials,
colours, and finishes for walls, floors, furniture, and decor elements, allowing for
personalized customization.
 DESIGN & MODELLING
Class Diagrams: These diagrams help in understanding the system's architecture and
facilitate the implementation of features like user management, design customization,
visualization, and AI recommendations.

Definition: A class diagram is an illustration of the relationships and source code

dependencies among classes in the Unified Modeling Language (UML). A class defines the
methods and variables in an object, which is a specific entity in a program or the unit of code
representing that entity.

Fig 1: Class Diagram representation

Sequence Diagram: These diagrams illustrate the interactions between different

system components over time to accomplish specific tasks.

Definition: A sequence diagram is a type of interaction diagram because it describes how

and in what order a group of objects works together. These diagrams are used by software
developers and business professionals to understand requirements for a new system or to
document an existing process.
Fig 2: Sequence Diagram representation

Use case Diagram: These diagrams depict the various ways users interact with the
system, highlighting the functional requirements and user goals.

Definition: It is a graphical depiction of a user's possible interactions with a system. A use

case diagram shows various use cases and different types of users the system has .The use
cases are represented by either circles or ellipses.

Fig 3: Use case Diagram representation

 MODEL BUILDING
To create a comprehensive AI-powered house design application, we can use a variety of
machine learning algorithms and data types. Here's an overview of how we can leverage
different algorithms to train models for generating design recommendations, visualizations,
and customizations.

Model Building
 Transfer Learning: Using pre-trained models like MobileNetV2, AI leverages
knowledge from existing models to improve performance and reduce training time.
 Custom Models: Custom AI models are built by adding new layers to pre-trained
networks to tailor them to specific tasks such as predicting design elements that match
user preferences.
 Other Algorithms: AI involves training various machine learning models (e.g.,
Random Forests, KNN) to handle different types of data and tasks.

Model selection
1. Convolutional Neural Networks (CNNs): For image-based data, CNNs are highly
effective. We'll use Transfer Learning with pre-trained models like MobileNetV2.
2. Recurrent Neural Networks (RNNs): For sequential data, such as time-series or
design sequences.
3. Decision Trees and Random Forests: For categorical and numerical data, these
algorithms can provide interpretable models.
4. K-Nearest Neighbours (KNN): For recommendation systems based on user
preferences.
5. Clustering Algorithms (e.g., K-Means): For grouping similar design elements or
user preferences.

Model Training
Train the models using the prepared data.

Model Evaluation
Evaluate all models based on accuracy, precision, recall, F1 score, etc., to select the best-
performing one for the final application.

Model Integration

Save the best model and integrate it into the application.

Project Architecture

The AI-Powered House Design Application follows a client-server architecture, with the
frontend and backend components interacting via RESTful APIs. The backend handles user
authentication, data storage, and communication with the AI module. The frontend provides
an interactive interface for users to customize and visualize house layouts, while the AI
module generates design recommendations based on user preferences.

Fig 4: Architecture for AI-Powered House Design

 CONCLUSION
In conclusion, the "Design Dream" project stands as a testament to the power of artificial
intelligence in revolutionizing the field of house design visualization and customization.
Through meticulous planning, innovative implementation, and user-centric design, we have
crafted a platform that empowers individuals to explore, conceptualize, and personalize their
dream homes like never before. In essence, "Design Dream" represents more than just a tool
for house design—it's a gateway to creativity, empowerment, and self-expression.
 REFERENCES
Data set reference
Smith,J.(2020).HousingPreferencesSurvey[Dataset].Kaggle.
https://www.kaggle.com/dataset/housing-preferences

Libraries
TensorFlow. (2015). Google. https://www.tensorflow.org/

PyTorch. (2016). Facebook. https://pytorch.org/

Scikit-learn:

Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M.,
Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., Passos, A., Cournapeau, D., Brucher,
M., Perrot, M., &Duchesnay, É. (2011). Scikit-learn: Machine Learning in Python. Journal of
Machine Learning Research, 12, 2825–2830. https://scikit-learn.org/stable/

Frameworks
Pallets Projects. (2010). Flask. https://flask.palletsprojects.com/

Docker. (2013). Docker. https://www.docker.com/

Twitter, Inc. (2011). Bootstrap. https://getbootstrap.com/ [Framework]