Related Experiment Video
Updated: Aug 5, 2026
08:08
Automated Measurement of Cryptococcal Species Polysaccharide Capsule and Cell Body
Published on: January 11, 2018
Image-based identification and DEA-based optimization modeling of antibiotic packaging using unsupervised learning
Phakdee Sukpornsawan1, Yutthapoom Meepradist1, Titinun Auamnoy1
1Faculty of Pharmaceutical Sciences, Burapha University, Chonburi, Thailand.
Plos One
|July 27, 2026
Summary
Accurate antibiotic package identification for medication safety is improved by combining image analysis and optical character recognition (OCR). This approach enhances pharmacy automation and dispensing systems.
Area of Science:
- Pharmaceutical Sciences
- Computer Science
- Image Analysis
Background:
- Medication safety relies on accurate identification of antibiotic packaging, crucial for pharmacy automation and dispensing.
- Advanced imaging and machine learning present innovative solutions for physical package recognition.
Purpose of the Study:
- To examine visual and textual features of antibiotic packages.
- To assess the relationship between these features and identification accuracy using unsupervised learning and efficiency analysis.
Main Methods:
- Analysis of 36 antibiotic formulations using binary imaging, entropy metrics, packaging area ratio (PAR), and optical character recognition (OCR).
- Application of K-means clustering for package segmentation and data envelopment analysis (DEA) for efficiency assessment.
Main Results:
- Identification of nine distinct image clusters.
- Association of mid-range entropy (7.1-7.5) and PAR (1.2-1.45) with higher identification consistency.
- Influence of OCR text confidence on identification outcomes; DEA identified efficient clusters.
Conclusions:
- Integration of image-derived metrics and OCR features facilitates automated antibiotic package identification.
- The developed framework offers a structured method for evaluating packaging characteristics in pharmacy workflows.
