摘要
类器官是一种结构和功能高度相似于人体内组织或器官的体外三维细胞聚集体。传统的类器官分析完全依赖对显微镜下捕捉到的类器官图像进行人工标记,这对其多维特征的量化是困难且耗时的。人工智能(AI)技术可自动地对大规模数据进行高通量分析,是更高效的类器官可视化分析模型。本文以计算机视觉领域中常用于类器官量化的技术为基础,对类器官可视化分析的三项基本任务,即图像分类、图像分割及图像跟踪展开综述。旨在总结AI在类器官筛选中的构建策略、应用前景及现存局限性,同时深入分析多任务协同下的多通道网络模型,最终为理解器官发育机制与疾病发展进程提供全新研究视角。
Organoids are three-dimensional cell aggregates that closely mimic the structure and function of human tissues or organs.Traditional organoid analysis relies entirely on manual labeling of microscopic images,which proves challenging and time-consuming for quantifying their multidimensional features.Artificial intelligence(AI)technology enables automated high-throughput analysis of large-scale data,offering a more efficient model for organoid visualization.This paper reviews three fundamental tasks in organoid visualization analysis—image classification,segmentation,and tracking—based on computer vision techniques commonly used for quantifying organoids.The study aims to summarize AI’s construction strategies,application prospects,and limitations in organoid screening,while analyzing multichannel network models with integrated tasks to provide new insights into organ development and disease progression.
作者
苗百卉
许建婷
丛宪玲
MIAO Baihui;XU Jianting;CONG Xianling(China-Japan Union Hospital of Jilin University,Changchun130033,China;Cancer Center,The First Norman Bethune Hospital of Jilin University,Changchun 130021,China)
出处
《中国实验诊断学》
2026年第1期135-139,共5页
Chinese Journal of Laboratory Diagnosis
关键词
人工智能
深度学习
类器官
可视化分析
高通量分析
Artificial intelligence
Deep learning
Organoids
Visualization
High-throughput analysis
