Search

Deep learning has substantially pushed forward remote sensing image change detection through extracting discriminative hierarchical features. However, as the increasingly high-resolution remote sensing images have abundant spatial details but limited spectral information, the use of conventional backbone networks would give rise to blurry boundaries between different semantics among hierarchical features. This explains why most false alarms in the final predictions distribute around change boundaries. To alleviate the problem, we pay attention to feature refinement and propose deep learning networks that deliver improved separability (ISNet). Our ISNet reaps the advantages from two strategies applied to refining bitemporal feature hierarchies: 1) margin maximization that clarifies the gap between changed and unchanged semantics and 2) targeted arrangement of attention mechanisms that direct the use of channel attention (CA) and spatial attention (SA) for highlighting semantic and positional information, respectively. Specifically, we insert CA modules into share-weighted backbone networks to facilitate semantic-specific feature extraction. The semantic boundaries in the extracted bitemporal hierarchical features are then clarified by margin maximization modules, followed by SA modules to enhance positional change responses. A top–down fusion pathway makes the final refined features cover multiscale representations and have strong separability for remote sensing image change detection. Extensive experimental evaluations demonstrate that our ISNet achieves state-of-the-art performance on the LEVIR-CD, SYSU-CD, and Season-Varying datasets in terms of overall accuracy (OA), Intersection-of-Union (IoU), and F1 score. Code is available at https://github.com/xingronaldo/ISNet. [ABSTRACT FROM AUTHOR]

Summary: Background: Liver fibrosis is the strongest histological risk factor for liver‐related complications and mortality in metabolic dysfunction‐associated fatty liver disease (MAFLD). Second harmonic generation/two‐photon excitation fluorescence (SHG/TPEF) is a powerful tool for label‐free two‐dimensional and three‐dimensional tissue visualisation that shows promise in liver fibrosis assessment. Aim: To investigate combining multi‐photon microscopy (MPM) and deep learning techniques to develop and validate a new automated quantitative histological classification tool, named AutoFibroNet (Automated Liver Fibrosis Grading Network), for accurately staging liver fibrosis in MAFLD. Methods: AutoFibroNet was developed in a training cohort that consisted of 203 Chinese adults with biopsy‐confirmed MAFLD. Three deep learning models (VGG16, ResNet34, and MobileNet V3) were used to train pre‐processed images and test data sets. Multi‐layer perceptrons were used to fuse data (deep learning features, clinical features, and manual features) to build a joint model. This model was then validated in two further independent cohorts. Results: AutoFibroNet showed good discrimination in the training set. For F0, F1, F2 and F3‐4 fibrosis stages, the area under the receiver operating characteristic curves (AUROC) of AutoFibroNet were 1.00, 0.99, 0.98 and 0.98. The AUROCs of F0, F1, F2 and F3‐4 fibrosis stages for AutoFibroNet in the two validation cohorts were 0.99, 0.83, 0.80 and 0.90 and 1.00, 0.83, 0.80 and 0.94, respectively, showing a good discriminatory ability in different cohorts. Conclusion: AutoFibroNet is an automated quantitative tool that accurately identifies histological stages of liver fibrosis in Chinese individuals with MAFLD. [ABSTRACT FROM AUTHOR]