Cell Anomaly Localisation using Structured Uncertainty Prediction Networks
Cell Anomaly Localisation using Structured Uncertainty Prediction Networks
复制标题
DOI：
--
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Boyko Vodenicharski;Samuel McDermott
中科院分区：
文献类型：
--
作者：
Boyko Vodenicharski;Samuel McDermott
本文提出了一种无监督的方法来异常检测在亮场或荧光细胞显微镜，我们的目标是定位疟疾寄生虫。这是通过构建描述健康细胞图像的生成模型（变分自编码器）来实现的，其中我们还对预测图像不确定性的结构进行建模，而不是在似然函数中假设像素独立性。这提供了一个“白化”的残差表示，其中生成模型的预期结构化错误被减少，但在训练分布中没有出现的独特结构，例如寄生虫被突出显示。我们采用最近发表的结构化不确定性预测网络方法来实现不确定性结构的可处理学习。在这里，残差协方差矩阵使用稀疏Cholesky参数化有效地逼近。我们证明，与对角高斯似然相比，我们提出的方法在检测真实和合成结构化图像扰动方面更有效。
This paper proposes an unsupervised approach to anomaly detection in bright-field or fluorescence cell microscopy, where our goal is to localise malaria parasites. This is achieved by building a generative model (a variational autoencoder) that describes healthy cell images, where we additionally model the structure of the predicted image uncertainty, rather than assuming pixelwise independence in the likelihood function. This provides a “whitened” residual representation, where the anticipated structured mistakes by the generative model are reduced, but distinctive structures that did not occur in the training distribution, e.g. parasites are highlighted. We employ the recently published Structured Uncertainty Prediction Networks approach to enable tractable learning of the uncertainty structure. Here, the residual covariance matrix is efficiently approximated using a sparse Cholesky parame-terisation. We demonstrate that our proposed approach is more effective for detecting real and synthetic structured image perturbations compared to diagonal Gaussian likelihoods.