Authors:
Kumudha Madesh, Kannammal Ayyasamy, Suresh Chockkan Chetty, Vinitha Thangavel, S. J. Vimal Aravintha, Raja Brahmendra Chowdary Veerepalli, Adithi Venkatakrishnan

Addresses:
Department of Computer Science and Engineering, Jayalakshmi Institute of Technology, Dharmapuri, Tamil Nadu, India. Department of Information Technology, Jayalakshmi Institute of Technology, Dharmapuri, Tamil Nadu, India. Department of Data Science, Indiana University Bloomington, Bloomington, Indiana, United States of America. Department of Computer Science, Texas State University, San Marcos, Texas, United States of America. Department of Information Technology and Management, University of Texas at Dallas, Richardson, Texas, United States of America.

Abstract:

The pharmaceutical sector is experiencing some serious problems in terms of drug mistakes and the emergence of fake medications. This study presents a powerful hybrid architecture suitable for detecting and identifying medicinal pills with high accuracy using a multi-stage deep learning framework. This research uses a dataset of 479 high-resolution data points that include different shapes, colors and pill imprints. The system combines Convolutional Neural Networks, which extract features and Transformer-based networks, which recognize sequences and patterns, resulting in improved accuracy across different lighting conditions. The software used includes Python-based deep learning libraries, specialized image augmentation frameworks and hardware-based processing units required to deliver real-time performance. The findings indicate that combining spatial and contextual information greatly reduces misidentification rates compared with single models. This mixed design offers a scalable, automated pharmacy check system and a patient-side safety app that ensure high fidelity between physical and electronic prescriptions.

Keywords: Pharmaceutical Recognition; Feature Extraction; Medication Safety; Pharmaceutical Interventions; Healthcare Technology; Computer Vision; Hybrid Learning; Computer Vision.

Received on: 19/06/2025, Revised on: 14/08/2025, Accepted on: 11/11/2025, Published on: 10/08/2026

DOI: 10.69888/FTSCS.2026.000721

FMDB Transactions on Sustainable Computing Systems, 2026 Vol. 4 No. 3, Pages: 187-195

• Views : 23
• Downloads : 10

Download PDF