🎬 Emotional Storyboard Generator for Filmmakers

An AI-powered web application that transforms film scripts into emotionally expressive storyboards, helping filmmakers, writers, and content creators previsualize their stories—without manual sketching or artistic expertise.

Developed as a final-year project (Jan 2025 - Apr 2025).

---

🧠 Overview

This tool uses Natural Language Processing (NLP) to analyze the emotional tone of each scene in a script and generate corresponding visuals using Stable Diffusion. It streamlines the pre-production process and accelerates visual storytelling.

---

✨ Key Features

• 🎭 Emotion Detection: Scene-wise emotion classification using fine-tuned Hugging Face Transformers
• 🖼️ AI-Generated Visuals: Uses Stable Diffusion to generate scene illustrations matching the detected emotion
• 📝 Script Upload: Secure file upload and backend processing
• 🎛️ Interactive Storyboard: Edit, rearrange, or regenerate individual frames
• 🚀 Scalable Deployment: Flask-based backend with deployment on Vercel
• 💾 Database Support: Flask-Migrate for versioned DB schema migrations

---

🧪 Tech Stack

Category	Technologies Used
Backend	Flask, Flask-Migrate, Python
AI/NLP	HuggingFace Transformers (e.g., BERT, RoBERTa), Emotion Datasets
Image Generation	Stable Diffusion
Frontend	HTML/CSS (optional React integration)
Deployment	Vercel, Gunicorn (optional), GitHub Actions (CI/CD)
Tools/Other	Cursor IDE, Git, Postman (for API testing)

---

📂 Project Structure

🚀 Getting Started (Local Setup)

Prerequisites

• Python 3.8+
• virtualenv or conda
• HuggingFace API Key (for model downloads)
• Vercel account (for deployment)
• GPU (optional for Stable Diffusion)

Installation

git clone https://github.com/your-username/emotional-storyboard.git
cd emotional-storyboard
python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows
pip install -r requirements.txt

Run Locally

flask db upgrade
flask run

Open your browser and go to: http://localhost:5000

🛠 How It Works

Upload your script in .txt format. The script is split scene-wise and passed through the emotion classifier. Detected emotions are used as prompts for Stable Diffusion to generate matching visuals. The storyboard is rendered in an interactive UI for further editing/export.

📸 Example

Input Script: INT. HOSPITAL ROOM - NIGHT She lies still. The monitor beeps faintly. Her father holds her hand, tears streaming. Detected Emotion: Sadness Generated Scene:

🔐 Security & Privacy

Uploaded scripts are stored securely and automatically deleted after processing (configurable). No data is shared with third-party services.

📚 Citation

If you use this project for academic or development purposes, please cite it as:

@project{reddy2025emotionalstoryboard,
  title     = {Emotional Storyboard Generator for Filmmakers},
  author    = {Puchakayala, Thirupathi Reddy},
  year      = {2025},
  month     = {April},
  howpublished = {\url{https://github.com/ThirupathiReddyPuchakayala/EMOTIONAL-STORY-BOARD-GENERATOR}},
  note      = {An AI-powered web application to convert film scripts into emotion-driven storyboards using NLP and Stable Diffusion.}
}