Pill Detect AI: An Integrated Machine Learning Web Platform for Automated Pill Recognition and Drug Interaction Analysis
Year: 2026
Type: conference-paper
Abstract: Medication errors arising from pill misidentification represent a significant and preventable source of patient harm. This paper presents PillDetect AI, a Flask-based web application that combines image-based pill recognition with an integrated drug interaction checking system. The system supports both image upload and manual descriptive input, returning pill identification results enriched with dosage information, pharmacological category, contraindications, and known drug-drug interactions. The interaction checker classifies drug pairs by severity and returns structured clinical recommendations. A RESTful JSON API is provided for integration with electronic health record systems. Testing against five NIH DailyMed pill profiles yielded 94% accuracy under manual input and 88% under image-based input using a ResNet-50 model stub.
Authors: Jaiswal, Vaishnavi, Bano, Tehseen, Burli, Yashodha Venkatesh, Mariam, Maseera, Ananthaswamy, Lakshmi
Language: English
Subfield: Health Informatics
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7162861585
FWCI:
Cites:
Cited by:
Related to: