five

Dataset for: Targeted multispectral filter array design for endoscopic cancer detection in the gastrointestinal tract

收藏
DataCite Commons2024-12-17 更新2024-07-13 收录
下载链接:
https://www.repository.cam.ac.uk/handle/1810/368220
下载链接
链接失效反馈
官方服务:
资源简介:
Description of methods used for collection/generation of data: Datasets were previously published and were collected as described in https://doi.org/10.1117/1.JBO.26.10.106002 and https://doi.org/10.1002/jbio.202100078). The folder “Datasets” contains curated spectral datasets used based on previous clinical trials for Oesophageal and Colon tissues these are read by the respective Python notebooks The folder “Optimization” includes the Python notebooks used to generate the optimal filter arrays Resulting variables saved to file: [classification method]_[tissue type]¬_varibles.npz ( eg. “kNN_Ose_variables.npz”) The saved variables include: for each number of bands N ‘bands_[N]' centre wavelengths of the selected bands (nM) 'bws_[N]' FWHM of bandwidths of the selected bands (nM) ' pattern_[N]' optimal arrangement of selected bands ' accu_[N]' spectral classification accuracy for optimal bands 'spatial_acc_[N]' spatial classification accuracy for optimal bands 'labels_pred_[N]' classification labels These variables were condensed in “Band Optimization Summary 23 July 23.xlsx” for easier interpretation “Band Optimization Summary 23 July 23.xlsx” is a summary of the optimized filters “Summary.ipynb” plots the figures used in the paper and SI All information is summarised in the readme.doc file.

数据采集与生成方法说明:本数据集已先期发表,其采集流程严格遵循文献https://doi.org/10.1117/1.JBO.26.10.106002与https://doi.org/10.1002/jbio.202100078中记载的方法。 "Datasets"文件夹内包含经专业整理校准的光谱数据集,该数据集源自针对食管与结肠组织的既往临床试验,相关Jupyter Notebook脚本可读取该文件夹内的全部数据。 "Optimization"文件夹内包含用于生成最优滤光片阵列的Jupyter Notebook脚本。生成的变量将保存至格式为`[分类方法]_[组织类型]_variables.npz`的文件中(示例:"kNN_Oes_variables.npz")。 保存的变量涵盖针对每个波段数N的如下内容: 1. `bands_[N]`:所选波段的中心波长(单位:纳米,nm) 2. `bws_[N]`:所选波段的半高全宽(Full Width at Half Maximum,FWHM)带宽(单位:纳米,nm) 3. `pattern_[N]`:所选波段的最优排布方式 4. `accu_[N]`:最优波段对应的光谱分类准确率 5. `spatial_acc_[N]`:最优波段对应的空间分类准确率 6. `labels_pred_[N]`:分类预测标签 上述变量已汇总至"Band Optimization Summary 23 July 23.xlsx"文件中,以方便快速解读;该文件为优化后滤光片的汇总文档。 "Summary.ipynb"可生成论文及补充材料(Supplementary Information,SI)中所用的全部图表。所有相关信息均已汇总至"readme.doc"文件中。
提供机构:
Apollo - University of Cambridge Repository
创建时间:
2023-07-26
二维码
社区交流群
二维码
科研交流群
商业服务