[论文解读] Not Color Blind: AI Predicts Racial Identity from Black and White Retinal Vessel Segmentations

Aaron S. Coyner, Praveer Singh+6 more|arXiv (Cornell University)|Sep 28, 2021

Retinal Imaging and Analysis参考文献 31被引用 4

一句话总结

本研究表明，人工智能能够通过深度学习从此前被认为不包含种族信息的黑白视网膜血管段图像中准确预测种族身份。即使去除了颜色信息，并对血管亮度和宽度进行归一化处理，卷积神经网络仍实现了近乎完美的种族预测（AUC-PR最高达0.995），揭示了即使在移除显性色素线索后，种族偏见仍可能存在于人工智能模型中。

ABSTRACT

Background: Artificial intelligence (AI) may demonstrate racial bias when skin or choroidal pigmentation is present in medical images. Recent studies have shown that convolutional neural networks (CNNs) can predict race from images that were not previously thought to contain race-specific features. We evaluate whether grayscale retinal vessel maps (RVMs) of patients screened for retinopathy of prematurity (ROP) contain race-specific features. Methods: 4095 retinal fundus images (RFIs) were collected from 245 Black and White infants. A U-Net generated RVMs from RFIs, which were subsequently thresholded, binarized, or skeletonized. To determine whether RVM differences between Black and White eyes were physiological, CNNs were trained to predict race from color RFIs, raw RVMs, and thresholded, binarized, or skeletonized RVMs. Area under the precision-recall curve (AUC-PR) was evaluated. Findings: CNNs predicted race from RFIs near perfectly (image-level AUC-PR: 0.999, subject-level AUC-PR: 1.000). Raw RVMs were almost as informative as color RFIs (image-level AUC-PR: 0.938, subject-level AUC-PR: 0.995). Ultimately, CNNs were able to detect whether RFIs or RVMs were from Black or White babies, regardless of whether images contained color, vessel segmentation brightness differences were nullified, or vessel segmentation widths were normalized. Interpretation: AI can detect race from grayscale RVMs that were not thought to contain racial information. Two potential explanations for these findings are that: retinal vessels physiologically differ between Black and White babies or the U-Net segments the retinal vasculature differently for various fundus pigmentations. Either way, the implications remain the same: AI algorithms have potential to demonstrate racial bias in practice, even when preliminary attempts to remove such information from the underlying images appear to be successful.

研究动机与目标

• 调查灰度视网膜血管图（RVMs）是否包含肉眼难以察觉的残留种族信息。
• 评估人工智能模型是否能在亮度、宽度和分割差异被归一化后，仍能从RVMs中检测种族。
• 评估视网膜血管的生理差异或分割伪影是否导致RVMs中的种族预测。
• 强调在图像预处理以去除明显种族标记后，种族偏见在人工智能驱动的医疗诊断中仍存在风险。
• 证明即使尝试消除颜色和色素特征，从RVMs中进行种族预测的准确性依然很高。

提出的方法

• 从245名黑人和白人婴儿中收集视网膜眼底图像（RFIs），共计4,095张图像。
• 使用U-Net模型从RFIs生成视网膜血管图（RVMs），随后通过阈值化、二值化和细化处理，生成多种RVM变体。
• 训练卷积神经网络（CNNs）以使用彩色RFIs、原始RVMs和处理后的RVMs（阈值化、二值化、细化）进行种族分类。
• 在图像级别和受试者级别均使用精确率-召回率曲线下面积（AUC-PR）评估性能。
• 为测试鲁棒性，在模型训练前对各组之间的亮度差异和血管宽度变化进行了归一化处理。
• 通过比较不同图像类型下的模型性能，分离出种族预测是否依赖于残留色素或血管模式的结构性差异。

实验结果

研究问题

• RQ1 人工智能模型能否从此前未被认为包含种族信息的灰度视网膜血管图中准确预测种族身份？
• RQ2 视网膜血管形态的残留差异或分割伪影在多大程度上导致RVMs中的种族预测？
• RQ3 在RVMs中对亮度和血管宽度进行归一化后，是否能消除人工智能预测种族的能力？
• RQ4 所观察到的种族预测能力是源于黑人与白人婴儿视网膜血管的生理差异，还是源于分割过程中的模型偏差？
• RQ5 人工智能能否从已预处理以去除颜色和色素线索的图像中检测种族？

主要发现

• 卷积神经网络（CNNs）在原始彩色视网膜眼底图像上实现了近乎完美的种族预测，图像级别AUC-PR为0.999，受试者级别AUC-PR为1.000。
• 仅使用原始视网膜血管图（RVMs）即可实现高精度的种族预测，图像级别AUC-PR为0.938，受试者级别AUC-PR为0.995。
• 即使在对血管亮度和宽度进行归一化后，CNNs仍保持强劲性能，表明种族预测并非依赖于这些视觉线索。
• 预测种族的能力在所有RVM变体中均持续存在——包括阈值化、二值化和细化版本——表明其对预处理具有高度鲁棒性。
• 研究结果表明，视网膜血管的生理差异或与眼底色素相关的分割伪影，可能共同导致RVMs中的种族检测。
• 本研究揭示，即使图像经过处理以去除明显种族标记，人工智能模型仍可能继承种族偏见，这在临床人工智能应用中构成重大风险。