Adaptive Segmentation Of LCT Images For Cervical Cell Based On GVF Snake Model

Early detection and treatment can significantly reduce the mortality rate of cervical cancer, so universal screening of cervical cancer is of great weight. Basically, the current inspection techniques is cervical smear screening technology, and cervical smear screening technology still adopts manual film-reading method. This way is inefficient and affected easily by subjective factors, would reduce the accuracy of screening. Thus, it is important and significative to develop the Computer-aided diagnosis system for screening of cervical cancer cell images, with machine analysis instead of manual film-reading, which can improve not only the efficiency but also the accuracy rate.Currently there are several cervical smear screening technology, and liquid-based cytology testing technology (LCT) is a new screening technology. It has advantages of production speed and high quality, so it is widely used in domestic hospitals, and represents a future trend. In this thesis, cervical LCT cell images is the objects of studying, I communicate with the hospital specialists and do a lot of experiments firstly, then extract the characteristics of cervical LCT cell image. According to these characteristics, I propose an adaptive segmentation algorithm based on GVF Snake model.Snake model is image segmentation algorithm based on parameter contour curve. It is a combination of low-level image information and high-level knowledge. Under the guidance of the high-level knowledge and drive by image energy, the curve drawn to the minimum energy location, that is the target edge. This algorithm is largely made up for shortcomings of traditional medical image segmentation algorithms, but it itself still has some weaknesses, such as small action scope and could not pull the curve to the cavity. Consequently, Xu proposes the GVF Snake model, which with the GVF field instead of the traditional Gaussian potential energy field. It is an effective solution to the problems of traditional Snake model and widely used in image segmentation field.In this paper, the basic theory of GVF Snake model has been studied, according to the characteristics of cervical LCT images, an adaptive segmentation algorithm based on GVF Snake model is proposed. Firstly, set the parameters of the model and segment, and then evaluate the segmentation result, determine if the result is good then stop, otherwise adjust the parameters and segment again. Lots of experimental results show that the algorithm can achieve sound effects.