Revitalize

Inspiration

As college students, we feel the pressures of schoolwork, internships, future careers, and general life weighing down at us on the same time. Sometimes, it becomes overwhelming. However, something we all enjoy is nature, which serves as a calming, tranquil, and rejuvenating environment. We wanted to incorporate learning and relaxing into one application.

What it does

This app allows a user to log any flowers or plants they have, or anything they scan. In an effort to both teach people about the cycle of life, mainly flowers and other types of pants (in a future update), and support the mental health of those in need, our product allows you to manage your own personal garden and watch as your hard work flourishes. Your virtual garden will be a reflection of your real garden, and you will be sent reminders to take care of your precious plants as our database of flowers will know what each flower needs and how to care for it. Over, time, you will amass a large garden and find peace of mind.

How we built it

By using a variety of python scripts, HTML, javascript, ruby, jekyll, and more, we put together a webpage and focused on comfort and ease of use. Lots of work went into creating a robust API to communicate with our database of plants which was created based off of a machine learning algorithm allowing you to identify flowers based on just a photo! Incredible work and ingenuity went into designing the webpage. This is truly a work of art.

Challenges we ran into

There were a number of challenges we ran into. There were troubles getting familiar with API calls and how to effectively communicate between the frontend and the backend. These issues take time to resolve, as there are a lot of moving pieces. In addition, There were some issues getting the model to train, as the dataset provided was not sorted nor labeled. This proved to be tedious work, as one wrong variable name could cause the entire dataset to be deleted.

Accomplishments that we're proud of

Overall, we are completely proud of our work. We trained a model which is able to classify flowers into one of 6 categories. Additionally we were able to manage several parts of the software development process after throwing ourselves into an ambitious project. There were many frustrating times during this project, but together we pulled through and were able to push out an amazing product. We are most proud of our API and Machine Learning algorithm built to manage and identity these flowers.

What we learned

As most hackathoners' do, we learned a lot in a lot of different areas. For example, we learned that if the size of the images are not all the same during the training of a model, the model will take this into account. In other words, if all dog images are larger and all cat images are small, the model may consider larger images to be dogs. This was interesting, and so we found that resizing images to the smallest image in the dataset (or removing the smallest if it is an outlier) must be done before training to avoid these situations. We all developed our understanding of API's, as they provide most of the communication in our program.

What's next for Revitalize

We have a lot of plans for ReVitalize, of which our next steps are fully connecting the backbone functionality of the program to the frontend to create a fully connected program. In addition to this, the model we currently use to predict plant class types is trained on a public dataset containing only 16 classes of flowers Moving forward, we would like to see the diversity of our model's ability expanded to a larger variety of plants.