CAJ | 학술논문

目的：通过研究肝癌计算机辅助诊断方法，实现自动识别肝癌病灶特征区域，达到辅助医生诊断的目的。方法：针对图像分割及识别等技术难点，在传统计算机辅助诊断的基础上，改进并提出新的图像预处理及图像分割方法；根据已有的大规模标注图像数据进行肝部病灶智能分类，从多方位对肝部病灶进行判断，以提高计算机辅助诊断的可用性。结果：通过试验可较好地对病灶区域进行分割及识别，初步实现了肝癌病灶的识别。结论：利用计算机辅助诊断方法，可敏锐地识别肝癌病灶的细微变化，为肝癌辅助诊断提供参考指标，提高肝部恶性肿瘤识别的准确性。
목적：통과연구간암계산궤보조진단방법，실현자동식별간암병조특정구역，체도보조의생진단적목적。방법：침대도상분할급식별등기술난점，재전통계산궤보조진단적기출상，개진병제출신적도상예처리급도상분할방법；근거이유적대규모표주도상수거진행간부병조지능분류，종다방위대간부병조진행판단，이제고계산궤보조진단적가용성。결과：통과시험가교호지대병조구역진행분할급식별，초보실현료간암병조적식별。결론：이용계산궤보조진단방법，가민예지식별간암병조적세미변화，위간암보조진단제공삼고지표，제고간부악성종류식별적준학성。
Objective:To investigate the computer auxiliary diagnosing method of liver cancer to realize the automatic recognition of hepatocellular carcinoma (HCC) lesions characteristic region so as to assist the doctor in diagnosing the tumor. Methods:Facing the technical difficulties in image segmentation and recognition, we put forward a new image preprocessing and segmentation method based on the data of traditional computer-aided diagnosis. Meanwhile, we made intelligent classification of liver lesions according to a large number of existing image annotation data to improve the usability of computer-aided diagnosis through the judgment of liver lesions in many aspects. Results: Better segmentation and recognition of the lesions were achieved to realize the identification of HCC lesions. Conclusion:The use of computer-aided diagnosis of HCC results in keen recognition of slight changes of lesions in patients. Furthermore, it can provide a reference for the auxiliary diagnosis of HCC. As a result, the diagnosing accuracy of liver malignant tumor can be improved combining with the doctor's clinical experience.