GAN · Random Forest · IoT

Design
floors,
intelligently.

Enter dimensions and room count — watch a trained GAN generate architecturally coherent floor plans. Powered by PyTorch, scikit-learn, and real-time IoT sensors.

GitHub
AI-generated · 256×256
100
Training Epochs
256px
Output Resolution
4
Core Modules
IoT
Real-Time Sensors
Arch-Ai-Tex
AI Floor Plan Generator
Scan QR for full project
Calculated Total Area: 1500.00 m² (≈ 16146 sq ft)
⏳ Loading model in background…
Rectangle
Square

Fetch ultrasonic readings one at a time from the ESP32 server and confirm whether it's Length or Breadth.

Arch-Ai-Bot
Architecture & Interior Design Assistant
Hey there! I'm your architecture assistant. Ask me anything about floor plans, room layouts, building design, or interior styles. 🏗️
01 — Features

Four modules.
One pipeline.

🏗️
GAN Floorplan Generator

Trained DCGAN synthesizes 256×256 floor plan images from noise + conditional inputs. Optional OpenCV denoiser and FCN-ResNet50 wall segmentation for crisp structural clarity.

📐
Optimized Layout Generator

Random Forest regression + heuristic algorithms produce spatially logical room arrangements tuned by plot dimensions, shape, and property type. Outputs 2D + interactive 3D.

📡
Real-Time Sensor Dashboard

HC-SR04 ultrasonic + PIR + IR sensors feed an Arduino Mega 2560 + ESP32. Live dimensions stream over Wi-Fi to auto-trigger GAN-based layout generation.

🤖
Arch-Ai-Bot

Built-in LLM chatbot (Llama 3.1 via Groq API) answers architecture and interior design questions with checklists and step-by-step guidance — always on in the sidebar.

02 — Process

From input
to layout.

1
Input Stage

Provide layout parameters manually — length, width, bedrooms — or let real-time sensor data do it automatically.

2
Generation Stage

The GAN generator synthesizes a floor plan. Simultaneously, the Optimized Layout module computes rule-based arrangements.

3
Segmentation

The generated plan passes through the segmentation model, producing a wall-highlighted binary mask for structural clarity.

4
Visualization

Results render with download options, denoising controls, and 3D modeling output.

03 — Stack

Built with
real tools.

Python 3.10
PyTorch
scikit-learn
Streamlit
OpenCV
Matplotlib
Plotly
Arduino Mega 2560
ESP32
HC-SR04
HC-SR501 PIR
FCN-ResNet50
Random Forest
DCGAN
Adam Optimizer
Groq API
Llama 3.1
C++
ARCH-AI-TEX

Try it now.

Enter your dimensions and watch the GAN generate a floor plan. No install needed.