A New Search Model for Liquid-based Cervical Cells Based on Auto-focus Microscope System

Abstract

Liquid-based cervical cytology test which one of the most important method of detecting cervical lesions has a good effect on diagnosis and prognosis of cervical carcinomas. Speed of manual scanning is slow, furthermore large quantities of works repeated due to the presence of visual fatigue that lead to misdiagnosis. Automatic screening method for pathological cells is proposed based on auto-focus microscope system combined advanced machine vision technology and image processing technology, convenient for doctors find suspicious cells. This way can improve the efficiency of medical workers and avoid misdiagnosis. The key technology of auto-focus microscope system is that scanning automatically of liquid-based cervical cell. An efficient and accurate focus method can find a clear image quickly, which is of great significance to the detection of abnormal cells. In this paper, we proposed a new research mode for auto-focus microscope system, based on the micrograph characteristics of liquid-based cervical cytology. Results show that the approach outperforms with high accuracy and comparable performance.