Increasing the Generalization of Supervised Fabric Anomaly Detection Methods to Unseen Fabrics.

Fabric anomaly detection (AD) tries to detect anomalies (i.e., defects) in fabrics, and fabric AD approaches are continuously improved with respect to their AD performance. However, developed solutions are known to generalize poorly to previously unseen fabrics, posing a crucial limitation to their applicability. Moreover, current research focuses on adapting converged models to previously unseen fabrics in a post hoc manner, rather than training models that generalize better in the first place. In our work, we explore this potential for the first time. Specifically, we propose that previously unseen fabrics can be regarded as shifts in the underlying data distribution. We therefore argue that factors which reportedly improve a model's resistance to distribution shifts should also improve the performance of supervised fabric AD methods on unseen fabrics. Hence, we assess the potential benefits of: (I) vicinal risk minimization (VRM) techniques adapted to the fabric AD use-case, (II) different loss functions, (III) ImageNet pre-training, (IV) dataset diversity, and (V) model architecture as well as model complexity. The subsequently performed large-scale analysis reveals that (I) only the VRM technique, AugMix, consistently improves performance on unseen fabrics; (II) hypersphere classifier outperforms other loss functions when combined with AugMix and (III) ImageNet pre-training, which is already beneficial on its own; (IV) increasing dataset diversity improves performance on unseen fabrics; and (V) architectures with better ImageNet performance also perform better on unseen fabrics, yet the same does not hold for more complex models. Notably, the results show that not all factors and techniques which reportedly improve a model's resistance to distribution shifts in natural images also improve the generalization of supervised fabric AD methods to unseen fabrics, demonstrating the necessity of our work. Additionally, we also assess whether the performance gains of models which generalize better propagate to post hoc adaptation methods and show this to be the case. Since no suitable fabric dataset was publicly available at the time of this work, we acquired our own fabric dataset, called OLP, as the basis for the above experiments. OLP consists of 38 complex, patterned fabrics, more than 6400 images in total, and is made publicly available.

The Medical Segmentation Decathlon.

International challenges have become the de facto standard for comparative assessment of image analysis algorithms. Although segmentation is the most widely investigated medical image processing task, the various challenges have been organized to focus only on specific clinical tasks. We organized the Medical Segmentation Decathlon (MSD)-a biomedical image analysis challenge, in which algorithms compete in a multitude of both tasks and modalities to investigate the hypothesis that a method capable of performing well on multiple tasks will generalize well to a previously unseen task and potentially outperform a custom-designed solution. MSD results confirmed this hypothesis, moreover, MSD winner continued generalizing well to a wide range of other clinical problems for the next two years. Three main conclusions can be drawn from this study: (1) state-of-the-art image segmentation algorithms generalize well when retrained on unseen tasks; (2) consistent algorithmic performance across multiple tasks is a strong surrogate of algorithmic generalizability; (3) the training of accurate AI segmentation models is now commoditized to scientists that are not versed in AI model training.

Deep-learning-based multi-class segmentation for automated, non-invasive routine assessment of human pluripotent stem cell culture status.

Human induced pluripotent stem cells (hiPSCs) are capable of differentiating into a variety of human tissue cells. They offer new opportunities for personalized medicine and drug screening. This requires large quantities of high quality hiPSCs, obtainable only via automated cultivation. One of the major requirements of an automated cultivation is a regular, non-invasive analysis of the cell condition, e.g. by whole-well microscopy. However, despite the urgency of this requirement, there are currently no automatic, image-processing-based solutions for multi-class routine quantification of this nature. This paper describes a method to fully automate the cell state recognition based on phase contrast microscopy and deep-learning. This approach can be used for in process control during an automated hiPSC cultivation. The U-Net based algorithm is capable of segmenting important parameters of hiPSC colony formation and can discriminate between the classes hiPSC colony, single cells, differentiated cells and dead cells. The model achieves more accurate results for the classes hiPSC colonies, differentiated cells, single hiPSCs and dead cells than visual estimation by a skilled expert. Furthermore, parameters for each hiPSC colony are derived directly from the classification result such as roundness, size, center of gravity and inclusions of other cells. These parameters provide localized information about the cell state and enable well based treatment of the cell culture in automated processes. Thus, the model can be exploited for routine, non-invasive image analysis during an automated hiPSC cultivation. This facilitates the generation of high quality hiPSC derived products for biomedical purposes.

The StemCellFactory: A Modular System Integration for Automated Generation and Expansion of Human Induced Pluripotent Stem Cells.

While human induced pluripotent stem cells (hiPSCs) provide novel prospects for disease-modeling, the high phenotypic variability seen across different lines demands usage of large hiPSC cohorts to decipher the impact of individual genetic variants. Thus, a much higher grade of parallelization, and throughput in the production of hiPSCs is needed, which can only be achieved by implementing automated solutions for cell reprogramming, and hiPSC expansion. Here, we describe the StemCellFactory, an automated, modular platform covering the entire process of hiPSC production, ranging from adult human fibroblast expansion, Sendai virus-based reprogramming to automated isolation, and parallel expansion of hiPSC clones. We have developed a feeder-free, Sendai virus-mediated reprogramming protocol suitable for cell culture processing via a robotic liquid handling unit that delivers footprint-free hiPSCs within 3 weeks with state-of-the-art efficiencies. Evolving hiPSC colonies are automatically detected, harvested, and clonally propagated in 24-well plates. In order to ensure high fidelity performance, we have implemented a high-speed microscope for in-process quality control, and image-based confluence measurements for automated dilution ratio calculation. This confluence-based splitting approach enables parallel, and individual expansion of hiPSCs in 24-well plates or scale-up in 6-well plates across at least 10 passages. Automatically expanded hiPSCs exhibit normal growth characteristics, and show sustained expression of the pluripotency associated stem cell marker TRA-1-60 over at least 5 weeks (10 passages). Our set-up enables automated, user-independent expansion of hiPSCs under fully defined conditions, and could be exploited to generate a large number of hiPSC lines for disease modeling, and drug screening at industrial scale, and quality.

Multiphase CT-based prediction of Child-Pugh classification: a machine learning approach.

BACKGROUND: To evaluate whether machine learning algorithms allow the prediction of Child-Pugh classification on clinical multiphase computed tomography (CT). METHODS: A total of 259 patients who underwent diagnostic abdominal CT (unenhanced, contrast-enhanced arterial, and venous phases) were included in this retrospective study. Child-Pugh scores were determined based on laboratory and clinical parameters. Linear regression (LR), Random Forest (RF), and convolutional neural network (CNN) algorithms were used to predict the Child-Pugh class. Their performances were compared to the prediction of experienced radiologists (ERs). Spearman correlation coefficients and accuracy were assessed for all predictive models. Additionally, a binary classification in low disease severity (Child-Pugh class A) and advanced disease severity (Child-Pugh class ≥ B) was performed. RESULTS: Eleven imaging features exhibited a significant correlation when adjusted for multiple comparisons with Child-Pugh class. Significant correlations between predicted and measured Child-Pugh classes were observed (ρLA = 0.35, ρRF = 0.32, ρCNN = 0.51, ρERs = 0.60; p < 0.001). Significantly better accuracies for the prediction of Child-Pugh classes versus no-information rate were found for CNN and ERs (p ≤ 0.034), not for LR and RF (p ≥ 0.384). For binary severity classification, the area under the curve at receiver operating characteristic analysis was significantly lower (p ≤ 0.042) for LR (0.71) and RF (0.69) than for CNN (0.80) and ERs (0.76), without significant differences between CNN and ERs (p = 0.144). CONCLUSIONS: The performance of a CNN in assessing Child-Pugh class based on multiphase abdominal CT images is comparable to that of ERs.