🖤 Shadow Monarch’s ASL Vision "Even in darkness, I see everything." — Sung Jin-Woo

This app is an ASL hand sign recognition tool inspired by the shadows. It uses a trained neural network to identify hand gestures from uploaded images. The current UI only supports drag-and-drop image uploads.

⚔ Prerequisites – Prepare Your Arsenal Before summoning predictions, equip your machine with these essential tools:

Download Node.js from here.

npm install npm run dev pip install -r requirements.txt uvicorn backend:app --reload --port 8000 pip install fastapi pip install uvicorn pip install keras pip install tensorflow pip install numpy pip install pandas pip install scikit-learn

🖐 How to Use — Wield the Power of Gesture Prepare your hand gesture as an image (28x28 grayscale preferred). Drag & drop the image into the drop zone of the app. The model will instantly predict the digit with precision.

📚 Dataset — Knowledge of the Shadows Our model is trained on the ancient Sign Language MNIST dataset, a collection of grayscale images representing American Sign Language (ASL) hand gestures. Each image is 28x28 pixels.

Classes: 24 letters (A–Y, excluding J and Z which require motion) Training Samples: 27,455 images Testing Samples: 7,172 images Format: Flattened 784-value grayscale images

🧠 Backend AI — Shadow Training Rituals The neural network model used for ASL classification is built using Keras and TensorFlow. Data exploration and preprocessing include:

Visualization via matplotlib and seaborn Data scaling using StandardScaler Dimensionality reduction with PCA and TSNE Model architecture includes Conv2D, MaxPooling, BatchNorm, Dropout, and L2 regularization

💡 Project Details — Echoes of the Void Frontend: React-based UI with a dark, themed interface. Backend: FastAPI serves model predictions through HTTP. Model: Trained convolutional neural network (CNN) built in Keras. Functionality: Users upload ASL digit images for recognition. Live Feedback: Predictions returned in real-time with animations. Theme: UI heavily inspired by Solo Leveling.

🌌 Final Words "The stronger you become, the more you can protect." — Sung Jin-Woo For more info click https://sign-language-analyzer-git-main-devams-projects-f85cb896.vercel.app.

THIS PROJECT DOSEN'T WORK ON PYTHON 3.13 (tensorflow is not available.)

NECESSARY LIBRARIES are given in src/pages/requirements.txt

This project isn't just code — it's a conduit for accessibility and style. May your gestures pierce through the digital void.