Dr.Pill

Instant identification of medical pills so they do not get wasted and those who do need them can use them correctly in a much more accessible fashion

Comment

Inspiration

According to the WHO, drug donation today is encountering serious waste in terms of both labor and medical products due to disorganized, massive amount of charity. 80% or donated drugs arrive unsolicited, unexpected, and most notably, unsorted. 62% come with labels in foreign languages that locals cannot decipher. Meanwhile, in other areas, drug cleanup fails to be completed appropriately and cause serious threats to the environment. Dr.Pill was then devised to simplify the medicinal identification process so people, ranging from those simply curious about the pills lying around in the cabinet to those in dire need of quick medical assortment, could utilize the current medical resources for efficiently and sustain-ably.

What it does

Dr.Phil uses image recognition to instantly identify medicinal pills, from over-the-counter to prescribed counterparts. Functional and consumption information after industry research was carefully selected among many from drugbank database, so contrary to traditional pill identifiers, Dr.Pill could provide much clearer and easier-to-follow insights. It also offers translation for international adoption, especially concerning the influence language barrier can build amidst drug donation.

How I built it

We created a server using node.js and separately wrote python scripts to use the machine learning APIs (Google OCR, IBM Watson, Translator), then we executed the scripts within the node.js server and rendered the results on frontend templates.

Challenges I ran into

Fetching data from drugbank and drugs.com was a challenging experience. Certain important drug-related information such as storage methods were inaccessibly as well.

Accomplishments that I'm proud of

We are proud to have given an attempt towards the life-sciences field for the first time. It was quite different and the topic was very rewarding.

What I learned

We fortified our experience on using computer vision APIs.

What's next for Dr.Pill

We hope to add advanced details such as storage details, actual dosages, and voice-interaction functionalities. Improvement in UX would be another viable integration with better readability of details.

Built With

Share this project:

Updates