Analysis and evaluation system for intestinal detection reports based on OCR

The gut microbiome testing report plays a crucial role in health assessment, disease diagnosis, nutritional adjustment, immune regulation, and health management. However, traditional paper reports are challenged to extract essential information and cannot conduct comprehensive comparisons and visual analyses of historical data. To tackle this issue, this study has been designed to develop a gut microbiome testing report analysis system based on OCR technology. The system utilizes classical machine learning and a small-sample classification method that integrates multi-scale graph features to extract crucial characters and feature indicator data from the detection reports. By training on three different neural network models, the system achieved classification accuracies of 97.82±0.16%, 99.02±0.23%, and 99.24±0.29%, respectively. The processed data can undergo multidimensional analysis, providing a foundation for nutritional health and resistance assessment. This system assists doctors in formulating more precise and effective treatment plans, ultimately enhancing patient treatment outcomes and quality of life.