Project Story
Inspiration
Our team decided to create a mobile app called "FlowerID" that utilizes deep learning to identify different types of flowers. We were inspired by the beauty of nature and the desire to learn about the different types of flowers that exist in the world.
How we built it
We used React Native to develop the app for Android devices, TensorFlow and PyTorch for the computer vision models, and Selenium and Beautiful Soup for web scraping flower descriptions. Our team trained two AI models: A YOLOV8 model that can recognize a flower in an image, which trained on custom dataset. And a MobileNetv2 model which is trained on Oxford102 Flowers Dataset. Hence, our app is capable of recognizing 102 different types of flowers.
Although we encountered some challenges along the way, such as optimizing the models for mobile devices and integrating them into the app, we ultimately succeeded in creating a functional and user-friendly app that can help people learn more about the flowers around them.
python
javascript
selenium
beautiful-soup
tensorflow
react-native
What We Learned
One of the biggest takeaways from this project was learning how to develop a mobile app using React Native. We also gained valuable experience working with deep learning frameworks like TensorFlow and PyTorch, as well as web scraping tools like Selenium and Beautiful Soup. Additionally, we learned about the importance of optimizing models for mobile devices, as they have limited resources compared to desktop computers.
Challenges We Faced
One of the biggest challenges we faced was integrating the AI models into the app in a way that optimized their performance on mobile devices. We also had some difficulty with training the models due to the large size of the Oxford Flowers 102 dataset. Finally, we didn't have enough time to implement the ability to take photos directly from the app, which meant users have to select a pre-existing photo from their gallery to identify the flower.
Overall, we're proud of what we've accomplished with FlowerID and we're excited to continue developing our skills in deep learning and mobile app development.
Log in or sign up for Devpost to join the conversation.