---

🧠 Image Similarity Search using Vision Transformer (ViT)

This project is a Streamlit web app that performs image similarity search using Vision Transformer (ViT) embeddings. Upload an image, and the model finds the most visually similar images from your dataset based on their ViT-generated embeddings.

🚀 Features

• Upload an image to search
• Compare it with a dataset of stored images
• Uses ViT (Vision Transformer) from TensorFlow Hub
• Computes cosine similarity between embeddings
• Returns top N similar images with similarity scores
• Easy-to-use Streamlit UI

🖼️ Demo

---

🧰 Tech Stack

• Python 🐍
• TensorFlow + Keras
• TensorFlow Hub
• NumPy
• Scikit-learn
• PIL (Pillow)
• Streamlit

---

📁 Project Structure

```bash . ├── app.py # Streamlit app ├── extract_embeddings.py # Script to precompute image embeddings ├── image_embeddings.npy # Saved ViT embeddings (generated) ├── images/ # Folder containing dataset images ├── requirements.txt # Dependencies └── README.md # This file

---

🔧 Setup Instructions

1. Clone the Repository

bash git clone https://github.com/your-username/image-similarity-vits.git cd image-similarity-vits

2. Install Dependencies

bash pip install -r requirements.txt

3. Add Your Images

Place all your dataset images inside the images/ folder (create if not exists).

4. Extract Embeddings (Run Once)

bash python extract_embeddings.py

This script uses ViT to generate and save embeddings in image_embeddings.npy.

5. Run the App

bash streamlit run app.py

---

🧠 How It Works

1. Loads ViT model from TensorFlow Hub
2. Precomputes embeddings for your dataset images
3. Accepts a query image through the UI
4. Extracts its embedding
5. Computes cosine similarity between the query image and dataset embeddings
6. Displays top N most similar images

---

⚠️ Notes

• Ensure your images are clear and relevant to your query domain.
• Preprocessing uses 224x224 image resizing to match ViT input.

---

📦 Requirements

tensorflow tensorflow-hub scikit-learn numpy pillow streamlit

Install using:

bash pip install -r requirements.txt

---

📸 Example Result

Query Image vs Top-5 Most Similar Images from Dataset

---

📄 License

This project is licensed under the MIT License. See LICENSE for details.

---

👨‍💻 Author

Nabeel Ali Khan Backend & AI Developer | Ethical Hacker GitHub: @nabeel03103n

---