Aiming at the problems of low efficiency, high false detection rate, and poor real-time performance of current industrial defect detection methods, this paper proposes an industrial defect detection method based on an expanded perceptual field and feature fusion for practical industrial applications. First, to improve the real-time performance of the network, the original network structure is enhanced by using depth-separable convolution to reduce the computation while ensuring the detection accuracy, and the critical information extraction from the feature map is enhanced by using MECA (More Efficient Channel Attention) attention to the detection network. To reduce the loss of small target detail information caused by the pooling operation, the ASPF (Atrous Spatial Pyramid Fast) module is constructed using dilate convolution with different void rates to extract more contextual information. Secondly, a new feature fusion method is proposed to fuse more detailed information by introducing a shallower feature map and using a dense multiscale weighting method to improve detection accuracy. Finally, in the model optimization process, the K-means++ algorithm is used to reconstruct the prediction frame to speed up the model’s convergence and verify the effectiveness of the combination of the Mish activation function and the SIoU loss function. The NEU-DET steel dataset and PCB dataset is used to test the effectiveness of the proposed model, and compared to the original YOLOv5s, our method in terms of mAP metrics by 6.5% and 1.4%, and in F1 by 5.74% and 1.33%, enabling fast detection of industrial surface defects to meet the needs of real industry.

Lung squamous cell carcinoma (LUSC) is a common histologic subtype of non-small cell lung cancer with a poor prognosis. RNA-binding proteins (RBPs) are key modulators in the posttranscriptional regulation and RBP alterations are commonly found in various cancer types. However, its roles in predicting the tumorigenesis and prognosis have not been identified in LUSC. To identify the roles of RBPs in the tumorigenesis and prognosis of LUSC, the RNA sequencing data of patients with LUSC were retrieved from The Cancer Genome Atlas (TCGA) databases. The differential expressed genes (DEGs) were evaluated and identified. The intersection of manually curated RBPs and tumorigenesis-related DEGs was filtered to the univariate Cox regression analysis. The intersection genes with prognostic value were defined as prognostic RNA-binding protein genes (PRBPGs). Based on them, the predicted model was constructed. Its accuracy was tested by the area under the curve (AUC) of the receiver operator characteristic curve and the risk score. In addition, to explore the key regulatory network, the relationship among PRBPGs, target RNA, and absolute quantification of 50 hallmarks of cancer was also identified by Pearson correlation analysis. A total of 311 genes were filtered as the intersection of 1542 manually curated RBPs and tumorigenesis-related DEGs and the results revealed 17 PRBPGs. Based on them, we constructed the predict model with a relatively high accuracy (AUC: 0.739). The Kaplan–Meier survival curve showed the significant prognostic value of risk score ( p < 0.001). Moreover, we uncovered the regulatory networks of PHF5A-TOMM22-oxidative phosphorylation, TLR3-CTSO inflammation-related pathway, SECISBP2L-targeted RNA (ADGRF5, TGFBR2, CD302, AC096921.2, AHCYL2, RPS6KA2, SLC34A2, and SFTPB) angiogenesis, and SECISBP2L-AKAP13 signaling (DNA repair, MTORC1 signaling, and MYC targets). The regulation mechanisms and cellular location of key PRBPGs were validated by assay for targeting accessible chromatin with high-throughput sequencing and single-cell RNA sequencing. Our study identifies PRBPGs as reliable indexes in predicting the tumorigenesis and prognosis of patients with LUSC and provides a well-applied model for predicting the overall survival for patients with LUSC. Besides, we also identified the regulatory network among PRBPGs, target RNA, and cancer gene sets in mediating the LUSC tumorigenesis.