HomeAboutProjects

Project

Mediscan

Compfest AI Innovation Challenge2023

Mediscan  Website
Mediscan, a participant in the Compfest AI Innovation Challenge, offers an app utilizing machine learning for doctor handwriting recognition and drug identification, along with a convenient pharmacy locator feature.

Team Members

  • Ikram Tauffiqul H
  • Mahendra Adi R
  • Novita Risdanti

Project Overview

Mediscan was developed as part of the Compfest AI Innovation Challenge, focusing on leveraging technology to improve healthcare services. Our app addresses the challenges of doctor handwriting recognition and drug identification through innovative machine learning techniques.

Key Features:

1. Doctor Handwriting Recognition

Doctor perception
Doctor perception

Utilizing YOLOv5, Mediscan trains a model to recognize real-world doctor handwriting, enabling accurate interpretation of prescriptions.

2. Drug Identification

Mediscan Drug Identification
Mediscan Drug Identification

Mediscan's second feature involves detecting and identifying drugs from their cover, providing users with detailed information about each drug.

3. Pharmacy Locator

Nearby Pharmacy
Nearby Pharmacy

Users can conveniently locate nearby pharmacies through Mediscan's integrated feature, facilitating easy access to medications.

Technical Details:

Drug Data Scraping

K24 web scrapping
K24 web scrapping

Python scripts with BeautifulSoup are used to scrape drug information from the K24Klik website, allowing for comprehensive drug information.

Deployment

Mediscan API Test
ML API Test

The entire application is containerized using Docker and deployed on Google Cloud Platform (GCP)