脑脊液细胞学AI分析系统的设计与实现  

Design and Implementation of AI System for Analysis of Cerebrospinal Fluid Cytology

作　　者：崔永刚 王梅[1] 陈德华[1] 潘乔[1] Cui Yonggang;Wang Mei;Chen Dehua;Pan Qiao(School of Computer Sciences and Technology,Donghua University,Shanghai 201620)

机构地区：[1]东华大学计算机科学与技术学院,上海201620

出　　处：《现代计算机》2022年第17期110-113,共4页Modern Computer

摘　　要：脑脊液细胞学检查是诊断脑膜炎、脑炎、梅毒等多类疾病的重要手段。随着AI技术的不断发展,使用AI进行脑脊液细胞分类及细胞学辅助诊断具有重要意义。本文从医院获得细胞涂片到生成检测结果的实际场景出发,设计并实现了一款前后端分离的脑脊液细胞学AI分析系统。基于深度神经网络目标检测和可解释算法,结合脑脊液细胞形态学特征提取,实现了细胞分类、结果可解释、细胞计数,以及辅助标注等功能。该系统有助于对各类型细胞进行精确量化的细胞学分析,可以大大节省医疗资源,对协助临床诊断以及治疗具有重要的应用价值。Analysis of cerebrospinal fluid cytology is the main means of diagnosing meningitis,encephalitis,syphilis and other diseases.As artificial intelligence continues to develop in the medical industry,it is a trend to use AI to help pathologists save themselves from monotonous repetitive work.In this paper,by using Faster R-CNN and S-Lime,combined with the characteristics of CSF cytology detection,we designed and implemented a CSF cytological AI analysis system for anterior-posterior separation,starting from the actual scene of obtaining microscope slides in hospitals and generating test results.The system realizes AI prediction,cell counting,auxiliary labeling and supplementary training,and is used to assist pathologists in cerebrospinal fluid cytology detection.More importantly,the system helps save medical resources and assist in clinical diagnosis.

关 键 词：计算机应用技术 目标检测 前后端分离 WEB开发 

分 类 号：R446.14[医药卫生—诊断学] TP18[医药卫生—临床医学] TP391.41[自动化与计算机技术—控制理论与控制工程]