ABSTRACT
Objective: Cervical smear screening for abnormal cells plays an important role in reducing the mortality of cervical cancer. However, all smears must be approved by the naked eyes using microscopes, and there has been a higher false-positive rate and false-negative rate. Methods: To solve this problem, the thin-layer cytology assistant screening system was developed. On the hardware side, the traditional computer-aided analysis system was improved; and on the software side, gray level transformation, edge detection, image mosaic, and some other image processing methods were introduced; and by the use of neural network, the system could realize the function of acquisition, mosaic, process, measure and recognition of the liquid-based cytology image. Results: This system was brought into the test of 150 thin-layer cytology technology(TCT) cervical cell samples, and the agreement between the results of computer-aided analysis and the manual analysis by pathology experts exceeded 90%. Conclusion: Both the traditional Pap smear and TCT preparation smear can be diagnosed by this system. As a result, the false-positive rate and false-negative rate of the cervical smear screening will be significantly reduced, meanwhile, work efficiency of pathologist can be greatly improved.