The advancement of computational technologies has propelled computer vision into multidisciplinary applications, notably in medical image processing. tissue semantic segmentation of pathological images is critical in ...
详细信息
The advancement of computational technologies has propelled computer vision into multidisciplinary applications, notably in medical image processing. tissue semantic segmentation of pathological images is critical in computational pathology. Deep learning models have shown promise in this task but often suffer from performance degradation when encountering out-of-distribution data due to variations in multi-center data distributions. To address this issue, we propose a dynamic single-domain generalization framework named tissue-SDG for tissue semantic segmentation. tissue-SDG incorporates dynamic adaptive data augmentation and multi-scale contrastive learning within a supervised framework to guide the model in learning domain-invariant features. Extensive experiments on our in-house CRC-GDPH-MS-tissueSeg dataset demonstrate the effectiveness of tissue-SDG, which outperforms existing methods. Specifically, tissue-SDG achieves an average Mean IoU of 79.64% across different test domains, exhibiting a remarkable improvement of at least 1% over baseline approaches. Here we show that our proposed techniques enhance the model's generalization ability and stability, making it more suitable for real-world applications. The source code is released at: https://***/liqian0926/tissue-SDG.
Accurate segmentation of tissue regions in hematoxylin and eosin (H&E)-stained histopathology images is an essential step in computational pathology. It is still a challenging task due to the large-scale distribut...
详细信息
Accurate segmentation of tissue regions in hematoxylin and eosin (H&E)-stained histopathology images is an essential step in computational pathology. It is still a challenging task due to the large-scale distribution, irregular morphological variations, and fuzzy boundaries between different tissues. semanticsegmentation methods with strong feature representative capability for extracting local fine-grained and global coarse-grained features are desired to tackle this challenge. In this paper, we propose a dual encoder network for tissue semantic segmentation of histopathology image, called DETisSeg, which aims to fully fuse the features of global context information by the Swin Transformer branch and local features by the convolutional neural network(CNN) branch. Particularly, we design an improved residual connection to recover the lost spatial information during the patch merging phase in the Swin Transformer branch. In addition, we employ a pyramid architecture decoder to generate a composite feature map with multiple scales. We perform experiments on two publicly available tissue semantic segmentation datasets, including the BCSS and LAUD-HistoSeg datasets, to evaluate the effectiveness of the proposed DETisSeg. Compared with eight state-of-the-art semanticsegmentation methods, the results show that the proposed DETisSeg achieves improved performance. Particularly, it outperforms Swin transformer on the mean Dice and IoU with improvements of 1.36% and 1.79% on BCSS dataset, 0.96% and 4.48% on LUAD-HistoSeg dataset, respectively.
暂无评论