Roles Of Expertise Knowledge In Medical Image Category Learning

Medical imaging is a kind of non-invasive image of the internal tissue of the body.Most medical imaging is very complex,so how can the doctor distinguish normal medical imaging from abnormal medical imaging? This is an important topic in the field of medical image understanding.Th research e study of visual perception learning shows that observers can distinguish different categories of images according to the statistical rules contained in the images.At the same time,the study of professional image understanding finds that expertise is also an important factor affecting image classification.So in the process of medical image classification,based on image statistical rules of visual perception and based on the concept of expertise learning is how to interact with each other,the novice is how to use the image of the accounting rules and the concept of expertise to image classification study,the effect of this kind of classification learning persistence,will this study is to explore the important scientific problems.This study included two experiments.The experimental materials were chest X-ray images,including normal images(healthy)and abnormal images(in the case of pulmonary nodules).The purpose of experiment 1 is to explore whether expertise of related concepts can promote visual perception classification learning based on image statistical rules.The independent variables were the expertise in medical imaging(expertise providing related concepts vs.expertise not providing related concepts)and the testing phase(pre-test,post-test).The 2 x 2 mixed design of experiment,comparison group have relevant expertise of the concept of accuracy,reaction time and eye movement index(IA dwell time,IA fixation count)in the learning process before and after the change trend of(pre-test,post-test)before and after the test,to explore whether to provide expertise of the concept of medical image classification learning for beginners.The results showed that the correct rate of classifying medical images into normal and abnormal after comparative learning was significantly improved regardless of whether expertise of related concepts was provided.More importantly,there was a significant interaction between the presence or absence of expertise and the test stage.The improvement in the accuracy of posttest in the group providing expertise was significantly better than that in the group without expertise.This result means that visual perception learning based on image statistical rules can improve the classification of medical images,and concept-based expertise can promote this classification learning.On this basis,thepurpose of the experiment 2 was to investigate the effect of classification learning persistence,experiment of 2 x 3 mixed design of experiment,the experiment of the independent variable is a expertise,provide the relevant concepts of expertise,no relevant concepts of expertise)and test phase(a pretest,immediate posttest and delayed 5)weeks after the test,the results of repeated experiment 1 found that the more important is to delay 5 weeks after the test results have been fairly and immediate posttest.This result implies that the effects of this classification learning are long-lasting.Based on the results of the two experiments,the following conclusions can be drawn: concept-based expertise can promote visual perception classification learning based on image statistical rules,and the effect of such learning will be lasting.