Student creates prescription-reading app that wins tech competition

(Dan Tri) - With just one click of a prescription on the phone, the application researched by a group of students will become a personal nurse, processing all information on the prescription for the user.

In a technology competition for students in Ho Chi Minh City, a research team from the University of Natural Sciences (Vietnam National University, Ho Chi Minh City) quickly made an impression with Megconize - an application used as a "personal nurse".

The research team won second prize in a technology competition for students in Ho Chi Minh City (Photo: Provided by the character).

Winning second prize in the competition, Bui Van Hung (22 years old, a member of the research team) confided that he and his family members often forgot to take their medicine or did not take it on time. This prevented their health from recovering in the best way.

Realizing the problem, Hung and his colleagues have hatched a project that can help patients easily track their prescriptions and remind them to take them on time.

"Users just need to take a photo or upload a prescription image to the app, the system will automatically process and enter information into the app such as diagnosis, prescription date, drug name, dosage and usage of each drug," Hung said.

At this time, the system will notify the user on the phone to remind them to take their medicine (Photo: Character provided).

Nguyen Nhat Cuong (22 years old, member) said that in about 3 seconds, the system will extract information from the prescription and create a separate medication schedule for the user. When it is time, the application will notify on the phone to remind them to take the medication.

To avoid forgetting and losing paper prescriptions, the application also saves the history of medication use. In addition, users can search for information about each drug through the search bar.

According to Nhat Cuong, if only making a version to submit as required by the lecturer, it would be very simple. The assignment only requires that with an available invoice, how can the application read the drug name on the prescription is passed. However, the group has the ambition to bring the product to support the real life of consumers.

It is known that the research team has 5 members, coming from many different majors, taking turns taking on separate tasks.

"The group has 3 people researching the prescription processing model, 1 person programming the application and 1 person responsible for managing the entire database. The most time-consuming part is building the data model, which took more than 3 months," Cuong said.

After taking a photo of the prescription, the system will process the diagnostic information and the name of the prescribed medication (Photo: Provided by the character).

One of the challenges the team had to overcome was how to extract information from images in the most optimal way. Sometimes prescriptions taken by users will be blurry, smudged, and not aligned, making it difficult for the system to recognize and extract information.

Not only that, the cost of buying a high-speed server is also an obstacle. "Free servers often process and extract information slowly. This will make users feel time-consuming and uncomfortable," Hung shared.

To solve the above problem, the team had to improve the information processing model many times. The team tested several models to sharpen, denoise and process text on photos, so that they were accurate in a short time.

"The team tested the system with more than 2,000 prescriptions in various forms from clinics and hospitals. All were done manually because there was no available data. Although it took a lot of time, it will help the application recognize a variety of prescription forms, not just within a certain internal unit," said Cuong.

Currently, the application has received many positive responses and has more than 100 downloads on the digital platform. In the near future, the team plans to launch a new version with the ability to analyze disease diagnoses on prescriptions and suggest appropriate health care regimens for each individual.

Dawn

Dantri.com.vn