Handwritten Character Recognition with CNN | MNIST & EMNIST Deep Learning Project | TensorFlow/Keras
🚀 In this video, I showcase my Handwritten Character Recognition Project developed during my internship at CodeAlpha.
Using Convolutional Neural Networks (CNNs) with TensorFlow/Keras, this model can classify handwritten digits (0–9) and letters from two benchmark datasets:
📊 Datasets Used:
MNIST (Digits 0–9) → Achieved 99% Accuracy ✅
EMNIST Balanced (Digits + Letters) → Achieved ~89% Accuracy 🔡
⚡ Project Highlights:
✔️ Data Preprocessing & Augmentation
✔️ Model Training & Evaluation
✔️ Confusion Matrix & Visualization
✔️ Techniques: Batch Normalization, Dropout, Early Stopping
✔️ Built & Tested on Google Colab GPU
🔗 GitHub Repository:
👉 https://github.com/Sohag016/CodeAlpha...
🎯 This project demonstrates the power of Deep Learning in Image Classification and is a great step toward real-world AI & Computer Vision applications.
If you found this helpful, don’t forget to Like 👍, Comment 💬, Share 🔄, and Subscribe 🔔 for more Machine Learning & AI projects!