The detection of strip steel surface defects is critical to ensure the quality of strip steel productions. At present, many deep-learning-based methods have been presented and achieve outstanding performance. However, most of these methods ignore the frequency information among defect areas, which plays an important role in defect detection. This paper proposes a deep learning method to further improve defects segmentation effect based on existing methods, called Low-Pass U-Net. Since most defects on strip steel are located in high frequency areas, we implement a low-pass filter before downsampling in encoder, which prevents aliasing and separates out high frequency information. The high frequency feature is transferred into decoder to assist segmentation. Following previous works, we propose adaptive variance gaussian low-pass layer to generate different filters according to each spatial location of feature map with fewer computing resource. Besides, to detect defects at significantly different scales, an improved Hypercolumn module is adopted at the end of decoder to upsample and fuse the feature maps in different resolutions, where Subpixel replaces the bilinear interpolation to refine the upsampled results. The proposed method is validated on practical datasets and achieves considerable performance improvement (with best Dice Coefficient 0.903), which demonstrates the effectiveness of Low-Pass U-Net. The introduction of AVGLFL layer results in 3% increase of Dice Coefficient in a comparative inference time, which achieves a balance in performance, inference time and complexity.
Background This study aims to propose a breast cancer prediction model for early diagnosis and prognosis management of breast cancer. Objective In order to explore the pathogenesis of breast cancer and develop accurate breast cancer screening and treatment methods, we have used machine‐learning technologies to conduct an in‐depth study of breast cancer genetic data to obtain new breast cancer signature and prognostic prediction models. Methods We explored an optimal cluster by unsupervised clustering methods with different expression genes (DEGs) between normal (n = 113) and tumour (n = 1,102) samples. Using least absolute shrinkage and selection operator (LASSO) regression, we selected four biomarkers to develop a predictive model by Cox regression method in the training set (n = 1,083) and validated its predictive accuracy and independence in the testing sets (n = 2,480). Then Gene Set Enrichment Analysis (GSEA) revealed enriched biological pathways in clusters. Finally, we constructed a nomogram including this signature and other significant risk factors to predict survival rates in patients. Results Four mRNAs (CD163L1, QPRT, NKAIN1 and TP53AIP1) between two clusters from 4,938 DEGs were identified, and then a four‐gene model (risk scores = 0.454*CD163L1–0.360*NKAIN1 + 0.581*QPRT + 0.788*TP53AIP1) was established to divide patients into high‐ and low‐risk group with significantly different prognosis (p < 0.0001) in the training set. Integrated analysis revealed dysregulated molecular processes including predominantly oncogenic signalling pathway, cell cycle and DNA repair in high‐risk group but enriched metabolism pathway in low‐risk group. In addition, this model had similar predictive value (HR >1.60; p < 0.05) in three independent validation sets, which could predict survival independently with more power compared with single clinical factor. In addition, the nomogram could predict the prognosis of breast cancer patients precisely in the training set and another three testing sets. Conclusion This model could predict prognosis of breast cancer patients precisely and independently, and provide evidence to make treatment decisions and design clinical trials.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.