Pokédex Classifier

Overview

This project is a Pokédex-inspired Pokémon classifier built using TensorFlow with EfficientNetB0 as the base model. The frontend is designed using PyQt6, replicating the UI/UX of a Pokédex. Users can upload an image of a Pokémon, and the classifier predicts its name. Additionally, Pokémon attributes are fetched using an open-source RESTful API.

Features

• Image Classification : Uses a TensorFlow model with EfficientNetB0 to classify Pokémon.
• Pokédex UI : Designed using PyQt6 with a background image resembling a Pokédex.
• File Upload : Users can select an image from their file system.
• Pokémon Attribute Fetching : Fetch Pokémon details via PokéAPI.
• Real-time Prediction : The model predicts and displays the Pokémon name along with its attributes.

Model and Dataset Details

• Framework : TensorFlow
• Base Model : EfficientNetB0
• Dataset : 1000 Pokemon Dataset

Future Improvements

• Optimizing the model by experimenting with advanced techniques such as fine-tuning, or incorporating more powerful architectures to increase prediction accuracy.
• Expand the model by incorporating additional Pokémon classes to enhance its classification capabilities.
• Design a better GUI for the application.