Automatic metallic surface defect inspection has received increased attention in relation to the quality control of industrial products. Metallic defect detection is usually performed against complex industrial scenarios, presenting an interesting but challenging problem. Traditional methods are based on image processing or shallow machine learning techniques, but these can only detect defects under specific detection conditions, such as obvious defect contours with strong contrast and low noise, at certain scales, or under specific illumination conditions. This paper discusses the automatic detection of metallic defects with a twofold procedure that accurately localizes and classifies defects appearing in input images captured from real industrial environments. A novel cascaded autoencoder (CASAE) architecture is designed for segmenting and localizing defects. The cascading network transforms the input defect image into a pixel-wise prediction mask based on semantic segmentation. The defect regions of segmented results are classified into their specific classes via a compact convolutional neural network (CNN). Metallic defects under various conditions can be successfully detected using an industrial dataset. The experimental results demonstrate that this method meets the robustness and accuracy requirements for metallic defect detection. Meanwhile, it can also be extended to other detection applications.

Objectives: Evidence of the relationshipbetween android fat mass and gynoid fat mass with the mortality prediction is still limited. Current study analyzed the NHANES database to investigate the relationship between android fat mass, gynoid fat mass and CVD, with all-cause mortality.Method: The study subjects were NHANES participants over 20 years old, two indicators of regional body composition, android fat and gynoid fat were measured by Dual Energy x-ray Absorptiometry (DEXA). The other various covariates data obtained from the NHANES questionnaire and laboratory measurements, including age, gender, education, race/ethnicity, uric acid, total serum cholesterol, albumin, Vitamin C, folate, alcohol drinking, smoking status, history of diabetes, and hypertension. Mortality status was ascertained from a linked mortality file prepared by the National Center for Health Statistics. The study population was divided quartiles based on the distribution of android fat mass and gynoid fat mass. The relationship between these two indicators with cardiovascular and all-cause mortality was investigated by using Cox regression. The covariates age, gender, smoking status, drinking status, history of diabetes, and history of hypertension were stratified.Results: In the fully adjusted model, Q3 had the lowest HR in android fat mass and gynoid fat mass. When examining the relationship between android fat mass and CVD mortality, current smokers and drinkers had the lowest CVD risk in Q2 [smoking: 0.21 (0.08, 0.52), drinking: 0.14 (0.04, 0.50)]. In diabetic patients, compared with Q1, other groups with increased android fat mass can significantly reduce the risk of CVD [Q4: 0.17 (0.04, 0.75), Q3: 0.18 (0.03, 1.09), Q2: 0.27 (0.09, 0.83)]. In ≥60 years old and female, the greater the gynoid fat mass, the smaller the HR of all-cause mortality [Q4 for ≥60 years old: 0.57 (0.33, 0.96), Q4 for female: 0.37 (0.23, 0.58)]. People <60 years old had a lower risk of all-cause mortality with gynoid fat mass in Q3 than those ≥60 years old [<60 years: 0.50 (0.27, 0.91), ≥60 years: 0.65 (0.45, 0.95)]. Among subjects without hypertension, the group with the largest android fat mass had the lowest risk of CVD mortality, and the group with the largest gynoid fat mass had the lowest risk of all-cause mortality [Android fat mass: 0.36 (0.16, 0.81), gynoid fat mass: 0.57 (0.39, 0.85)].Conclusion: Moderate android fat mass and gynoid fat mass (Q3) had the most protective effect. Smokers and drinkers need to control their body fat. Being too thin is harmful to people with diabetes. Increased gynoid fat mass is a protective factor for all-cause mortality in older adults and females. Young people's gynoid fat mass is more protective in the moderate range than older people's. If no high blood pressure exists, people with more android and gynoid fat mass have a lower risk of CVD or all-cause mortality.

Scratches are one of the most common defects in industrial manufacturing. Weak scratches in the industrial environment have an ambiguous edge, low contrast, large span, and unfixed shape, which bring difficulty for automatic defect detection. Recently, many existing visual inspection methods based on deep learning cannot completely and effectively inspect industrial weak scratches due to the lack of discriminative features and sufficient spatial detail. In this article, a novel DeepScratchNet is proposed for automatic weak scratch detection by aggregating rich multidimensional feature for scratch representation. To obtain rich features, a pretrained ResNet block as a feature extractor is proposed in this article. To highlight features of scratch and weaken the noise, an attention feature fusion block (AFB) is proposed, which densely fuses high-level semantic features with low-level detail features using dual-attention mechanism. Due to the long span and connectivity of the weak scratches, a context fusion block (CFB) is proposed to learn the complete context. To further improve the scratch segmentation performance, the auxiliary loss is integrated into the proposed network. The proposed DeepScratchNet outperforms the traditional and other state-of-the-art deep learning-based methods on three given real-world industrial data sets with mIoU over 0.8005, 0.812, and 0.9286. The experimental results demonstrate that DeepScratchNet achieves good generalization capabilities.

The detection of conductive particle images is an important part of the chip on glass (COG) process and can be used to ensure the performance of electrical connections. The segmentation of conductive particles is essential but a difficult task, since the scale and edge of the conductive particles on the chip and the imaging effect are different. In recent years, methods based on deep learning have become the representative method of image segmentation. However, the currently existing methods cannot fully consider the characteristics of conductive particles and have high model complexity. In this article, a multi-frequency feature learning-based convolutional neural network (CNN) is proposed. The entire network structure consists of a basic U-Net module and multi-frequency module (MFM), which are used to enhance multi-frequency feature fusion of conductive particles and accelerate network training. At the same time, for the feature of particle shape, an active contour without edge (ACWE) loss function is designed to extract the fine contour feature of particles. Experimental results on three datasets show the superiority of the proposed method over the major existing mainstream methods with respect to the three performance indicators: recall, precision, and Intersection-over-Union (IoU).

Spermatogenesis begins when cell fate-committed prospermatogonia migrate to the basement membrane and initiate spermatogenesis in response to retinoic acid (RA) in the neonatal testis. The underlying cellular and molecular mechanisms in this process are not fully understood. Here, we report findings on the involvement of a cancer/testis antigen, PRAMEL1, in the initiation and maintenance of spermatogenesis. By analyzing mouse models with either global or conditional Pramel1 inactivation, we found that PRAMEL1 regulates the RA responsiveness of the subtypes of prospermatogonia in the neonatal testis, and affects their homing process during the initiation of spermatogenesis. Pramel1 deficiency led to increased fecundity in juvenile males and decreased fecundity in mature males. In addition, Pramel1 deficiency resulted in a regional Sertoli cell-only phenotype during the first round of spermatogenesis, which was rescued by administration of the RA inhibitor WIN18,446, suggesting that PRAMEL1 functions as an inhibitor of RA signaling in germ cells. Overall, our findings suggest that PRAMEL1 fine-tunes RA signaling, playing a crucial role in the proper establishment of the first and subsequent rounds of spermatogenesis.