A Preliminary Research Of The Quantitative Palm-print Characteristics With Relation To Gastric Cancer

Numerous studies have demonstrated that there are some relationships between dermotoglyphics and disease, such as some congenital hereditary diseases including some kinds of tumors. Many genetic or congenital diseases are showed certain palm-plait grain texture of the differences. At present, the medicine on dermatoglyph study is still dependent on qualitative research and experts subjective experience judgment, lack of all-round quantitative, digital characterization and evaluation system, it makes research is hard to obtain more rapidly on the relationships between dermatoglyph and disease.The main aim of study to shows that the relationship of the disease patients who have special dermatoglyph and he is liable to disease. Analysis through clinical trials, the unity of the palm-plait grain characteristic index system with quantitative and digitization can convert come subjective description of palm-plait grain characteristic into more accurate mathematical model, and quantitative description the palm-plait grain associated with happening of the disease by means of this model that we establishment. This not only provides a unified evaluation method and tool for medical dermotoglyphics researchers, but also played a certain role in promoting to the development of clinical dermatoglyph.The task in this paper, we should be convert the subjective and qualitative description of the characteristic indexes on palm-plait grain into quantitative analysis, and realize digital said through the establishment of quantitative palm-plait grain characteristic index system. In experiment, we used the gastric cancer patients'palm-plait grain characteristics as clinical date sources, and analyze the differences with the quantitatively method of the palm-plait grain between the patients and controls, in order to discover the inherent law from mathematical perspective reflect relationships between palm-plait grain and gastric cancer. Classify the gastric cancer patients and controls into two types by palm-plait grain characteristics through further using Fisher discriminance and SVM method, comparison of classification results; we obtained the classifier model that it is suitable for the palm-plait grain characteristic index system.This thesis main job as follows:1. To construct a data platform for collection palm-plait grain which can real-time acquire and process image and then obtain the features data.2. To analysis and extract some palm-plait grain related to the gastric cancer based on the existing expert experience and knowledge of traditional Chinese medicine.3. To quantitatively described the palm-plait grain in the feature set.4. Take some statistical methods such as chi-square, T test, normality analysis, correlation analysis and Fisher discriminant, etc. we do a lot of statistics analysis on the each palm-plait grain and contrastive analysis between the gastric cancer and controls to obtain 14 palm-plait grain features and establish quantitative feature indices. 5. To do two-class Fisher discrimination analysis in order to confirm the validity of palm-plait grain features being chosen and take these results prove that the features can be used to distinguish the gastric cancer and the controls, then to do Logistic analysis in order to establish the equation which relates to the gastric cancer susceptibility and palm plait grain features.6. To design SVM to classify the patients and the controls according to the palm-plait grain features, and compare the effect on determining the best classifier.Conclusion:First, there are some statistically significant differences between the gastric cancer patients and control groups in palm-plait grain, which provides a reliable statistically assistance for establishing a palm-plait grain characteristic index system and for classifying between the patients and controls.Secondly, choose the correct mathematical tools and appropriate variables based on the analysis of the palm-plait grain's types and characteristics with the scientific methods, can use mathematical methods to expression and description features, and then establish the mathematical model.Thirdly, the model identification technology and use correct statistics method can proved that correct description and quantify for the palm-plait grain can preliminary forecast gastric cancer, and to achieve early screening and prevention.Lastly, gastric cancer patients and controls are classified according to the palm-plait grain features indices by SVM, results showed that SVM is better than others.