用于算法配准分割识别的显微图像与OCT三维数据集

主要面向多模态成像融合、配准与智能分割识别算法的研究与模型训练需求建设。该数据集基于自研OCT-显微镜联合成像系统获取，结合高分辨率显微成像与光学相干断层成像技术，对眼科组织仿体、离体眼及微结构标准靶进行同步采集，生成空间对应的二维显微图像与三维OCT体数据。通过亚像素级空间标定与自动配准算法，实现显微图像与OCT体数据的精确对齐。数据内容包括原始显微图像序列、三维OCT体数据、结构层标签、针尖与组织分割掩膜以及配准标注文件，可用于多模态分割、特征融合及语义识别算法的训练与评估。数据总量约为276 MB。该数据集为OCT与显微成像在术中导航、组织识别及智能辅助决策中的算法验证提供了高质量基础支撑与标准化实验平台。

This dataset is constructed to address the research and model training needs of multimodal imaging fusion, registration, and intelligent segmentation and recognition algorithms. It is acquired based on an independently developed OCT-microscopy combined imaging system, which combines high-resolution microscopy imaging and optical coherence tomography (OCT) technology. The system synchronously collects ophthalmic tissue phantoms, ex vivo eyes, and microstructure standard targets, generating spatially corresponding two-dimensional (2D) microscopy images and three-dimensional (3D) OCT volumetric data. Sub-pixel-level spatial calibration and automatic registration algorithms are employed to achieve precise alignment between the microscopy images and OCT volumetric data. The dataset includes raw microscopy image sequences, 3D OCT volumetric data, structural layer annotations, needle and tissue segmentation masks, and registration annotation files, which can be utilized for the training and evaluation of multimodal segmentation, feature fusion, and semantic recognition algorithms. The total size of the dataset is approximately 276 MB. This dataset provides high-quality basic support and a standardized experimental platform for algorithm validation of OCT and microscopy imaging in applications including intraoperative navigation, tissue recognition, and intelligent auxiliary decision-making.

搜集汇总

背景与挑战

背景概述

该数据集专为多模态成像融合、配准与智能分割识别算法的研究与模型训练而构建。它基于自研的OCT-显微镜联合成像系统，提供了精确对齐的高分辨率二维显微图像与三维OCT体数据，并包含结构层标签、分割掩膜等多种标注文件，数据总量约为276 MB，为相关算法的验证与评估提供了高质量的基础数据支撑。