A Method Of Research Of General Classification Methods On Image

Cervical cancer is one of the most common female reproductive tract malignant tumor, in the process of liquid early cervical cytology screening, We need methods are roughly divided a whole to the image content. This method provides basis behind the cell classification model.Based on the medical technology co., LTD. Nanchang silver dell early cervical cancer screening system platform, after the thin liquid layer cervical cells image acquisition mode this paper put forward a kind of based on feature modeling and the law of gray histogram parameter automatic detection method of intelligent analysis, then achieve the rapid general cervical classified image content. By the late early cervical cancer screening system platform for different specimens. The main content of this paper include: A. The law of the histogram analysis and selection of feature points and feature parameters. B. The specimens of normal and abnormal histogram regularity of specimen analysis and modeling. C. The method of automatically detecting the concept and application of histogram parameters.The law of the histogram analysis and selection of feature parameters is based on the analysis of the gray histogram of cell research compiled histogram of four general rules, then according to the histogram of summed up the four general rules, choose can reflect the characteristics of gray histogram feature points. The characteristic parameters is determined by six feature points in the end. This paper has chosen six feature points with seven characteristic parameters as the histogram model reaction.Specimens of specimens of normal and abnormal regularity of histogram analysis and model is established in this paper. Specimens of normal histogram model is by means of the description on the cellular pathology of combining histogram additive. First analyze the general rule of single cell histogram, thus finishing the gray-level histogram model of normal specimens, and the effective of the model validation. Abnormal histogram regularity of specimen analysis and model building is divided into three parts- specimens of glandular cell, blood specimens of pollution and inflammation. Analyze abnormal pathological characteristics of specimens, and the embodiment of these characteristics on the histogram analysis, and combining with the general law of histogram, modeling of abnormal samples, finally to verify this abnormal samples of different models.Histogram parameter automatic detection method is based on the specimens of histogram analysis and histogram feature points on the automatic detection and characteristic parameters in the process, the problems are analyzed and described, indicating the solution, finally through online experiments verify its effectiveness.This paper presents a method of automatic measurement based on histogram parameters.it is applied to the liquid based cervical cell image intelligent screening system, achieves roughly divided the integrity of image content.