Improving Visual Defect Detection and Localization in Industrial Thermal Images Using Autoencoders.

Reliable functionality in anomaly detection in thermal image datasets is crucial for defect detection of industrial products. Nevertheless, achieving reliable functionality is challenging, especially when datasets are image sequences captured during equipment runtime with a smooth transition from healthy to defective images. This causes contamination of healthy training data with defective samples. Anomaly detection methods based on autoencoders are susceptible to a slight violation of a clean training dataset and lead to challenging threshold determination for sample classification. This paper indicates that combining anomaly scores leads to better threshold determination that effectively separates healthy and defective data. Our research results show that our approach helps to overcome these challenges. The autoencoder models in our research are trained with healthy images optimizing two loss functions: mean squared error (MSE) and structural similarity index measure (SSIM). Anomaly score outputs are used for classification. Three anomaly scores are applied: MSE, SSIM, and kernel density estimation (KDE). The proposed method is trained and tested on the 32 × 32-sized thermal images, including one contaminated dataset. The model achieved the following average accuracies across the datasets: MSE, 95.33%; SSIM, 88.37%; and KDE, 92.81%. Using a combination of anomaly scores could assist in solving a low classification accuracy. The use of KDE improves performance when healthy training data are contaminated. The MSE+ and SSIM+ methods, as well as two parameters to control quantitative anomaly localization using SSIM, are introduced.

Inference of tissue relative proportions of the breast epithelial cell types luminal progenitor, basal, and luminal mature.

Single-cell analysis has revolutionised genomic science in recent years. However, due to cost and other practical considerations, single-cell analyses are impossible for studies based on medium or large patient cohorts. For example, a single-cell analysis usually costs thousands of euros for one tissue sample from one volunteer, meaning that typical studies using single-cell analyses are based on very few individuals. While single-cell genomic data can be used to examine the phenotype of individual cells, cell-type deconvolution methods are required to track the quantities of these cells in bulk-tissue genomic data. Hormone receptor negative breast cancers are highly aggressive, and are thought to originate from a subtype of epithelial cells called the luminal progenitor. In this paper, we show how to quantify the number of luminal progenitor cells as well as other epithelial subtypes in breast tissue samples using DNA and RNA based measurements. We find elevated levels of cells which resemble these hormone receptor negative luminal progenitor cells in breast tumour biopsies of hormone receptor negative cancers, as well as in healthy breast tissue samples from BRCA1 (FANCS) mutation carriers. We also find that breast tumours from carriers of heterozygous mutations in non-BRCA Fanconi Anaemia pathway genes are much more likely to be hormone receptor negative. These findings have implications for understanding hormone receptor negative breast cancers, and for breast cancer screening in carriers of heterozygous mutations of Fanconi Anaemia pathway genes.