The paper presents a novel deep learning-based approach, MIPC-Net, for precise boundary segmentation in medical images. The key components of the approach are:
Mutual Inclusion of Position and Channel Attention (MIPC) module: This module enhances the focus on channel information when extracting position features and vice versa, mimicking radiologists' working patterns to improve the precision of boundary segmentation.
GL-MIPC-Residue: This global residual connection enhances the integration of the encoder and decoder by filtering out invalid information and restoring the most effective information lost during the feature extraction process, improving the restoration of medical images.
The proposed MIPC-Net model is evaluated on three publicly accessible datasets: Synapse, ISIC2018-Task, and Segpc. The ablation study shows that each module contributes to improving the quality of segmentation results. Furthermore, with the assistance of both modules, MIPC-Net outperforms state-of-the-art methods across all metrics on the benchmark datasets, notably achieving a 2.23mm reduction in Hausdorff Distance on the Synapse dataset, strongly evidencing the model's enhanced capability for precise image boundary segmentation.
Egy másik nyelvre
a forrásanyagból
arxiv.org
Mélyebb kérdések