🧠 MNIST Digit Recognition App

A sleek and interactive AI-powered digit recognition app built using Python, Streamlit, OpenCV, and Scikit-Learn.

---

🚀 Features

✅ Upload, Draw, or Use Camera – Three ways to input a digit
✅ Camera Feed – Capture image using your computer’s camera
✅ Interactive Charts – Visualize model confidence levels
✅ Adaptive Thresholding – Improves digit recognition accuracy
✅ Smooth UI & Animations – Modern sidebar, CSS enhancements, and loading effects

---

📂 How It Works

1️⃣ User Chooses an Input Method:

• Upload Image – Upload a handwritten digit image
• Draw Image – Use an interactive canvas to draw a digit
• Camera Image – Capture a digit using the webcam

2️⃣ Image Processing:

• Converts the image to grayscale
• Applies adaptive thresholding for better contrast
• Resizes to 28×28 pixels (MNIST format)
• Inverts colors (black digit on white background)
• Normalizes and flattens the image for model input

3️⃣ Prediction:

• The Extra Trees Classifier predicts the digit
• Displays confidence levels in an interactive bar chart

4️⃣ Results:

• The predicted digit is displayed with a success message
• A probability distribution chart visualizes confidence scores

---

🛠️ Technologies Used

• Python (Core logic)
• Streamlit (Web app framework)
• OpenCV (Image processing)
• Scikit-Learn (Machine learning models)
• Altair (Data visualization)
• Pillow (Image manipulation)
• Joblib (Model loading)

---

💻 How to Run

1️⃣ Clone the Repository

git clone https://github.com/Mrcolgate2024/NBI_python_delkurs2_kunskapskontroll

cd NBI_python_delkurs2_kunskapskontroll

2️⃣ Install Dependencies

pip install -r requirements.txt

3️⃣ Run the App

streamlit run DigitRecog_App.py

---

🎯 Future Improvements

🔹 Add more machine learning models for comparison

🔹 Enhance UI with animations and custom themes

🔹 Deploy the app online for public access - digitsense.streamlit.app")

---

🚀 Built by Bjorn R Axelsson