Poke Dexscan - College Project

Poke Dexscan is a modern Pokédex web application built with Flask and powered by a TensorFlow machine learning model. Upload Pokémon images to identify them using a trained CNN, and explore details like types, locations, and evolutions through an interactive scan interface and Pokémon list. Features a sleek red/blue aesthetic with pixel font and drag-and-drop upload.

Features

• Image Upload & Classification: Drag-and-drop Pokémon images with live preview; identifies using a TensorFlow CNN model.
• Interactive UI: Flask backend with HTML/CSS/JS frontend; responsive design with Pokéball loading animation.
• Pokémon List: Grid view of all Pokémon with modal details (description, type, location, evolutions).
• Professional Design: Red Pokédex frame, blue screen, pixel font ('Press Start 2P'), and hover effects.
• Robust Processing: Supports any image format (PNG, JPG, AVIF) with background removal via rembg for accurate predictions.

Project Structure

Poke Dexscan/
├── app.py                  # Flask backend with routes and model loading
├── pokedex_data.py         # Pokémon data (descriptions, types, etc.)
├── train_cnn_model.py      # Training script for CNN model
├── requirements.txt        # Dependencies
├── pokemon_cnn_model.h5    # Trained TensorFlow model
├── pokemon_classes.npy     # Model class names
├── frontend/               # Static frontend files
│   ├── index.html          # Home page (scan interface)
│   ├── list.html           # Pokémon list page
│   ├── styles.css          # Global styling with Pokédex theme
│   ├── script.js           # JS for upload and scan
│   ├── script_list.js      # JS for list and modal
│   └── assets/
│       └── pokeball.png    # Loading animation
└── README.md

Setup and Run

1. Clone the Repository
   git clone https://github.com/rubel286/Poke-Dexscan.git
   cd Poke-Dexscan

2. Install Dependencies
   pip install -r requirements.txt

3. Add Model Files
   Train pokemon_cnn_model.h5 and pokemon_classes.npy using train_cnn_model.py.
   Place them in the root directory.

4. Run the Application
   python app.py

Access at: http://127.0.0.1:5000

• Home: Upload/scan images
• /list: Browse Pokémon

Usage

• Scan: Drag an image or click to upload; view prediction with confidence score.
• List: Click a Pokémon card for modal details.
• Supported Formats: PNG, JPG, AVIF (via wand/rembg).