Data Models for Dataset Drift Controls in Machine Learning With Optical Images - Datasets

This dataset accompanies the paper titled Data Models for Dataset Drift Controls in Machine Learning with Images that appeared in the Transactions on Machine Learning Research https://openreview.net/forum?id=I4IkGmgFJz @article{ oala2023data, title={Data Models for Dataset Drift Controls in Machine Learning With Optical Images}, author={Luis Oala and Marco Aversa and Gabriel Nobis and Kurt Willis and Yoan Neuenschwander and Mich{\`e}le Buck and Christian Matek and Jerome Extermann and Enrico Pomarico and Wojciech Samek and Roderick Murray-Smith and Christoph Clausen and Bruno Sanguinetti}, journal={Transactions on Machine Learning Research}, issn={2835-8856}, year={2023}, url={https://openreview.net/forum?id=I4IkGmgFJz}, note={} } We make available two datasets. Raw-Microscopy: 940 raw bright-field microscopy images of human blood smear slides for leukocyte classification (microscopy/images/raw_scale100) with corresponding labels (microscopy/labels). 5,640 variations measured at six additional different intensities (microscopy/images/raw_scale001-raw_scale0075) 11,280 images of the raw sensor data processed through twelve different pipelines (microscopy/images/processed_views) Raw-Drone: 548 raw drone camera images for car segmentation (drone/images_tiles_256/raw_scale100) with corresponding binary segmentation mask (drone/masks_tiles_256). The images and the masks are cropped from 12 raw drone camera images (drone/images_full/raw_scale100) and 12 masks (drone/masks_full) of size 3648 by 5472. 3,288 variations measured at six additional different intensities (drone/images_tiles_256/raw_scale001-raw_scale075). 6,576 images of the raw sensor data processed through twelve different pipelines (drone/images_tiles_256/processed_views). Detailed datasheets for the two datasets can be found in the appendices of the TMLR paper. The code repository for this project can be found at https://github.com/aiaudit-org/raw2logit本数据集配套发表于《机器学习研究汇刊》（Transactions on Machine Learning Research）的论文，相关论文链接为https://openreview.net/forum?id=I4IkGmgFJz。对应文献条目如下： @article{ oala2023data, title={面向光学图像机器学习的数据集漂移控制数据模型}, author={Luis Oala、Marco Aversa、Gabriel Nobis、Kurt Willis、Yoan Neuenschwander、Michèle Buck、Christian Matek、Jerome Extermann、Enrico Pomarico、Wojciech Samek、Roderick Murray-Smith、Christoph Clausen、Bruno Sanguinetti}, journal={《机器学习研究汇刊》（Transactions on Machine Learning Research)}, issn={2835-8856}, year={2023}, url={https://openreview.net/forum?id=I4IkGmgFJz}, note={} } 本项目公开两类数据集： 1. Raw-显微镜成像数据集（Raw-Microscopy）：包含940张用于白细胞分类的人体血涂片明场显微镜原始图像（存储路径：microscopy/images/raw_scale100），配套对应标注文件（存储路径：microscopy/labels）；另有6组不同成像强度下生成的5640张变体图像（存储路径：microscopy/images/raw_scale001至raw_scale0075）；以及经12种不同处理管线得到的原始传感器数据图像共11280张（存储路径：microscopy/images/processed_views）。 2. Raw-无人机成像数据集（Raw-Drone）：包含548张用于车辆分割的无人机相机原始图像（存储路径：drone/images_tiles_256/raw_scale100），配套对应二值分割掩码（存储路径：drone/masks_tiles_256）。上述图像与掩码均由12张原始尺寸为3648×5472的无人机原始相机图像（存储路径：drone/images_full/raw_scale100）及12张对应掩码（存储路径：drone/masks_full）裁剪得到；另有6组不同成像强度下生成的3288张变体图像（存储路径：drone/images_tiles_256/raw_scale001至raw_scale075）；以及经12种不同处理管线得到的原始传感器数据图像共6576张（存储路径：drone/images_tiles_256/processed_views）。 两类数据集的详细技术手册可参见该《机器学习研究汇刊》论文的附录部分。本项目的代码仓库可访问：https://github.com/aiaudit-org/raw2logit