Team CHRYSTAL - MedEx

MedEx revolutionises Singapore's healthcare sector by utilising AI to simplify access to medical research articles, extract information from them, and compress them into infographics for convenience.

Comment

Inspiration

We were inspired by the presence of existing products which utilise AI to facilitate the extraction and compression of information from legal documents to ensure smoother workflows of legal processes. Upon studying various impacts arising from the use of such products, we found that we could implement similar systems for the healthcare sector, to ensure smoother workflows for healthcare professionals.

What it does

MedEx utilises search engine optimisation to facilitate access to the latest relevant articles related to the user’s search query. It also utilises AI-based abstract segmentation and natural language processing to extract and compress information from articles in the PDF format into infographics almost instantly.

How we built it

We built a workable website using html5 and css, and utilised Python and Flask to run our AI-based algorithms that are capable of extracting information from research articles in the form of PDF files, and also of compressing such information into infographics for presentation purposes.

Challenges we ran into

Lack of background experience in the implementation of AI and related APIs in our code, including abstract segmentation and natural language processing, prior to iNTUition v9.0

Accomplishments that we're proud of

We were able to devise and implement AI and related APIs in our code despite being in a high-pressure setting, i.e. 24 hours of solution prototyping and development

What we learned

Strong teamwork and team camaraderie amongst our group of three individual registrants helped ensure our team's success in developing our successful prototype of our product which could address our chosen problem statement (sponsored by MSD) directly.

What's next for Team CHRYSTAL - MedEx

We aim to implement AI-based smart search filters on a larger number and wider variety of medical research articles, with average accuracy of search results exceeding 99%. On top of that, we also intend to implement AI-based smart extraction and compression of information in research articles into infographics with a single click, even without the need to upload PDF files.

Built With

Share this project:

Updates