Дослідження питання використання фото упаковки ліків з метою автоматизації введення інформації в мобільний застосунок для обліку лікарських засобів у побуті

Abstract

This study investigates the use of multimodal large language models for automated extraction of structured medication data from images of pharmaceutical packaging. Addressing the limitations of manual data entry in existing mobile medication management applications, the research explores the capability of MLLMs to perform context-aware information extraction directly from photographs. A specialized dataset containing annotated images of medicines was developed to evaluate model performance across individual fields and tabular components. The results demonstrate that MLLM-based approaches enable reliable and semantically consistent data extraction. The proposed method can be integrated into mobile applications to automate medication entry and improve user experience in home medicine cabinet management systems