Inspiration
We wanted to create a fun, interactive way to explore and learn about real-world animals, something that feels like a game but teaches you something, too. Inspired by the nostalgic magic of the Pokédex and our love of nature, we imagined a device that could turn everyday animal encounters into moments of discovery. Whether you're a kid on a nature walk or a seasoned birdwatcher, Animadex brings a sense of wonder to spotting wildlife.
What it does
Animadex is a handheld device that lets users "catch" real animals by taking their photo. It uses machine learning to identify the species, displays a pixel-art version of the animal, and shares a brief description. Each identified animal is added to your personal Animadex collection—like a real-world Pokédex. It's designed to make nature exploration fun, educational, and collectible.
How we built it
We used an ESP32 microcontroller paired with an OV7670 camera to capture images. A TFT SPI screen displays the user interface, including the animal description and pixel-art graphic. Image classification is done using a machine learning model optimized for the ESP32. Animal data and captured entries are stored locally in flash memory. We also designed a simple UI and collection system to browse through previously discovered animals.
Challenges we ran into
- Getting the OV7670 camera to work reliably with the ESP32’s I2S interface was tough due to timing and memory constraints.
- Rendering images and UI elements in real-time on a low-power microcontroller required optimization.
- Creating or converting animal images into low-resolution pixel art that was both recognizable and memory-efficient.
- Making the model lightweight enough for real-time inference on the ESP32 without cloud processing.
Accomplishments that we're proud of
- Built a fully functional handheld device that identifies animals and displays pixel art in real time.
- Successfully integrated camera capture, ML inference, and display output in a single compact system.
- Designed a fun and accessible user experience for both kids and adults.
- Turned a nostalgic idea into a working, educational tool.
What we learned
- How to interface camera modules with microcontrollers using I2S.
- Efficiently managing memory and rendering on low-power hardware.
- Optimizing ML models for embedded systems.
- The value of blending education and play in tech design.
What's next for AniDex
- Add offline voice feedback to describe animals aloud.
- Support for more animal types, including insects and aquatic life.
- Include GPS tagging so users can see where they found each animal.
- Implement the function so that users' parents can see their location.
- Design a mobile companion app to sync and share Animadex collections.
- Build a custom 3D-printed case to make the device more rugged and portable.
Built With
- c
- esp32-cam
- raspberry-pi
- tft-screen
Log in or sign up for Devpost to join the conversation.