神经网络在卵巢浆液性囊腺肿瘤细胞核形态分析中的应用

目的：本研究将计算机技术与病理学专家的实际经验相结合，在采用图像处理技术对医学图像进行处理的基础上，应用人工神经网络分析卵巢浆液性囊腺肿瘤细胞核形态特征，并比较径向基函数网络与自组织网络两种算法。方法：对每个良性，交界性，或恶性卵巢浆液性囊腺肿瘤细胞及正常卵巢上皮组织细胞分别提取五个形态参数，用两种神经网络进行分类识别。结果：径向基函数训练步数少，但准确率没有自组织网络高。结论：应用神经网络进行细胞识别。在医学科研以及临床诊断方面有重要的现实意义和广阔的应用前景。

Application of Neural Network in Morphometric Study of Ovarian Serous Cystadenomatous Tumor

Objective： Computer technology is combined with practical experiences of pathology experts. On the basis of processing techniques of medicine images,netwal network is used to recognize the morphometrical features of ovarian serous cystadenomatous tumor using neural network, contrast two algorithms- radial-basis function and ad-hoc network. Materials and Methods： After the 5 morphometry parameters of cellualr nuclear were obtained from every serous cystadenoma cell,borderline serous cystadenoma cell, serous cystadenocarcinoma cell, classify them by two neural networks. Results： Radial-Basis Function has less Wain steps and more accuracy than ad-hoc network. Conclusions： Neural network is used to recognize cells images and shows its significant value in clinical diagnosis and medical research at present and in the future.