[Practical Application of Intelligent Vision Measurement System Based on Deep Learning].

Researchers

Journal

Modalities

Models

Abstract

To comprehensively assess the true visual function of clinical dry eye patients and the comprehensive impact of blinking characteristics on functional vision of the human eye, an intelligent vision measurement system has been designed and developed to detect and analyze blinks from the side. The system employs deep learning keypoint recognition technology to analyze eyelid features from a lateral perspective. It presents the data of identified key points for the upper and lower eyelids in a line chart format and annotates the trough of each blink. By setting benchmark values, the system automatically calculates the proportion of complete and incomplete blinks in the tested individuals. The results indicate that the system is stable in performance and accurate in measurement, successfully achieving the anticipated design objectives. It thereby provides reliable technical support for future clinical applications.