🎨 Text-to-Image Synthesis with DC-GAN: A Deep Learning Approach ⚡

🖌️ Transform text prompts into stunning visuals using a Deep Convolutional GAN (DC-GAN). This full-stack project combines NLP and generative models for creative image synthesis, complete with user interactions and social features.

🚀 Key Features

🧑💻 User Experience

• 📱 Responsive UI: Works flawlessly on mobile/desktop
• 🔒 Secure login/signup flow
• 🖼️ Dynamic Gallery:
  • ⬆️ Upload images (PNG/JPG)
  • ⬇️ Download with 1-click
  • ❤️ Like & 💬 Comment system
  • 🌐 Social sharing (Twitter/Facebook/WhatsApp)

🧠 AI & Backend

• 🌀 DC-GAN Model: Trained on 60k Cifar-100 dataset samples
• 🔄 Smart Sorting:
  • 🕒 Recent | 📅 Oldest | 🏆 Most liked
  • 🔍 Search by tags/descriptions

🛠️ Tech Stack

Frontend

Backend

🛠 Setup & Installation

1. Clone the repository

git clone https://github.com/Somepalli-Venkatesh/text2image

2. Install dependencies for the root

npm install

3. Start the frontend

cd frontend
npm install
npm run dev

4. Set up the backend

cd backend
pip install -r requirements.txt

5. Start the backend

python app_cim.py

Project Structure

text2Image/
├── client/               # React + Vite frontend application
│   ├── public/           # Static assets
│   └── src/              # Components, pages, styles
├── backend/               # Flask backend 
│   ├── static/gallery            
│   ├── app.py        
│   ├── bert.py          
│   ├── cifargenerator.h5          
│   ├── cifardiscriminator.h5     
│   ├── dcgan.py 
│   ├── image_gen.py  
│   └── requirements.txt          
└── README.md             # Project overview and setup

💾 Dataset

• Source: CIFAR-100 (60,000 images, 100 classes)
• Preprocessing:
  • Resize to 64×64
  • Normalize to [–1, +1]
  • Text prompts encoded with pre-trained BERT
• Augmentation:
  • Random flips, rotations, color jitter

🎓 Learning Resources

• GANs in Action (book) by Jakub Langr & Vladimir Bok
• Original DCGAN paper: “Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks” by Radford et al. (2015)
• Text-to-Image Survey: https://arxiv.org/abs/2008.03187
• TensorFlow GAN Tutorial: https://www.tensorflow.org/tutorials/generative/dcgan

📬 Contact Us

👤 Name	📧 Email Address
Venkatesh Someplli	venkateshsomepalli0@gmail.com
Tumati Manohar	manohartumati569@gmail.com
Yetukuri Venkata Kusuma	yvenkatakusuma2005@gmail.com
Tupakula Keethi	tkeerthi039@gmail.com