基于人工智能的近红外荧光分子断层成像数据集

荧光分子断层成像（FMT）可以获取探针的三维在体分布实现准确的肿瘤检测。然而，简化的辐射传输方程和复杂的逆问题会导致重建误差。基于本数据集，提出了一种基于神经网络的FMT重建方法，以降低定位误差、提高形态恢复度。仿真和活体实验证明，该方法提高了FMT的重建性能，促进了神经网络在光学成像研究中的应用。

Fluorescence Molecular Tomography (FMT) can obtain the 3D in vivo distribution of probes to achieve accurate tumor detection. However, simplified radiative transfer equations and complex inverse problems often lead to reconstruction errors. Based on this dataset, a neural network-based FMT reconstruction method is proposed to reduce localization errors and improve morphological recovery performance. Both simulation and in vivo experiments have verified that the proposed method enhances the reconstruction performance of FMT and promotes the application of neural networks in optical imaging research.

搜集汇总

背景与挑战

背景概述

该数据集专注于近红外荧光分子断层成像（FMT），旨在通过人工智能技术改进肿瘤检测中的三维探针分布重建。它提供了一种基于神经网络的FMT重建方法，以降低定位误差并提升形态恢复度，仿真和活体实验验证了其性能提升，促进了神经网络在光学成像领域的应用。