自适应脉冲神经膜系统的计算性能及其在胆石病自动识别中的应用研究

胆囊癌患者一般伴随有胆结石，两者具有较高的合并发病率，这使得胆石病的识别和诊断是胆囊癌预防的关键所在。基于脑启发计算模型的胆石病自动识别方法是智慧医疗框架下的重要研究课题之一，也是计算机科学和医学交叉研究的重要领域之一。脉冲神经膜系统是受人脑神经系统启发得到的计算模型，具有计算并行度高、易读易扩展、容错性强等特点。本项目研究自适应脉冲神经膜系统的计算性能及其在胆石病自动识别中的应用，研究内容包括：①构建拓扑结构和激发行为自适应调整的脉冲神经膜系统；②从计算能力、计算效率和计算复杂性三个方面研究自适应脉冲神经膜系统的计算性能；③结合胆石病例的B超和CT图像，发展基于自适应脉冲神经膜系统的胆石自动识别方法；④在胆石病临床筛查中应用验证，对模型进行反馈校正。预期研究成果可为脉冲神经膜系统提供新的动态演化工作模式、丰富系统的计算理论，还可在智慧医疗框架下为胆石病的自动识别和临床筛查提供新方法。

Cholecystolithiasis is usually accompanied by gallstone, which makes the identification and diagnosis of gallstone disease to be one of the key problems to the prevention of gallbladder cancer. Brain-inspired computing models have provide plenty of computation methods in intelligent medicine, particular in automatic recognition of gallstone disease. Spiking neural P systems are a novel class of brain-inspired parallel and distributed computing models, having features of computing in parallel, easily reading and extending, and high fault tolerance. Self-adaptive feature of spiking neural P systems is a hot and hard research topic in membrane computing, but hasn’t been fully studied. In this project, we plan to investigate self-adaptive spiking neural P systems, including the computational properties and their applications. There are 4 main research contents: ① design of spiking neural P systems with self-adaptive topological structure and self-adaptive spiking behavior; ② investigate the computational power, computational efficiency and computational complexity of the constructed self-adaptive spiking neural P systems; ③ design learning algorithms for self-adaptive spiking neural P system to process cholelithiasis patient data; ④ develop automatic recognition method for cholelithiasis disease based on spiking neural P systems and used in clinical diagnose. This study will provide new dynamic evolution strategy and computational theories for spiking neural P systems, and also provide novel automatic recognition method for cholelithiasis disease under the framework of intelligent clinic.