MediBox

The ML model in the system fills in the blank automatically.
Users can choose to scan medication box or add photo from their gallery.
My Purchase page allows people to add more medications
Users can check their medication progress and information on this page
Users can check their intake time and dose. If not taken on time, alarm will be triggered.
Please use this image as an example if you were to try our project out.

Inspiration

As medical research and achievements have greatly progressed in the recent decades, medication has become a common part of people's daily lives. Despite the beneficial outcome brought by medications, there can be a challenge remembering to take the correct dose at the right time. While some medical instructions are complicated, it would take time to figure out one's dose. Most people have experience about forgetting to take their medications or not taking on time. This is a phenomenon known as “medication non-adherence”

What it does

Our product, MediBox can read medication instructions and set alarms for reminding purposes. In this way, the patient will not have to manage any concerns about the intake of their medicines. In the meantime, we offer our user vouchers for helping us build our NLP library. As users grow, our NLP models will be more powerful in recognition and will be saving users tremendous amounts of time compared to manually entering the dose and intake time.

Value Proposition: For people on medications who tend to forget their intake time and indicated dose, our APP, MediBox, is a collective tool that reads instructions and notifies people to take the right dose at the right time.

How we built it

We used Flutter to develop our cross platform applications for Android, iOS, and so on. For the machine learning models, we applied open source package : https://github.com/tesseract-ocr/tesseract. We also used Firebase as our backend.

Challenges we ran into

Our biggest challenge is that we are working remotely from each other. Under this circumstance, we were unable to sit face-to-face and code together. Instead, we spent a lot of time merging our code and explaining what we have integrated. Another challenge we faced was the prefered programming languages of our team members were quite different. Some of the members had to compromise and quick learnt another programming language.

Accomplishments that we're proud of

We are the most proud that we have practiced responsible innovation. In this Hackathon, we present a product for understanding and supporting the public aimed at ‘responsible innovation’. MediBox saves users' time and makes their lives easier. We aim for the overall good of the whole society, caring for everyone especially those in unfavoured positions. We depend less on the users' skills on the technologies, but more automation service for elders or different abled people.

What we learned

We have practiced Agile during the development of our prototype. In this process, we polished our idea with a massive amount of discussion. We learned soft skills such as active listening, time-management, creative thinking, conflict resolution, and so on. The most important thing we have gained from this experience is our friendship and connections. Technically, we have learnt how to implement Technically, we have learnt how to integrate Machine Learning Model on the device end. This is the first time we have ever tried to do so, and we have enjoyed the learning journey very much.

What's next for MediBox

Allow users to order prescription or non-prescription medication with pharmacies before arrival.
Obtain a prescription from an online certified GP.
Send notification for medicine restock.
Extend to a fully functional platform: In the future, we will have mongodb for storing pictures, pytorch for training models, neo4j for relationship chain analysis, and data analysis such as spark in the future, as well as the background of the entire number of steps in the cloud, docker kubernetes, load balancing Nginx, elastic search for search, and graphql for unified api.