Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Table 1.

Novelty clarification of the proposed framework (high-level).

More »

Table 1 Expand

Fig 1.

The overall framework consists of an encoder driven by Boundary-augmented Hybrid Attention (BAHA) and a decoder guided by multi-scale shape priors, achieving joint modeling of lesion boundary features, cross-scale structural information, and global semantic context.

The decoder further integrates a feature pyramid structure, a gradient extraction module, and a multi-layer perceptron to produce high-precision skin lesion segmentation results with enhanced boundary refinement and improved shape consistency.

More »

Fig 1 Expand

Fig 2.

The structure of the Boundary-augmented Hybrid Attention (BAHA) module is illustrated, where the feature branch and gradient branch are first encoded using 1D CNNs and then processed by local self-attention and convolutional gating units to extract semantically relevant local contextual information.

The gradient features generate a boundary mask through Laplacian and EdgeConv operations, which guides channel and spatial reweighting, and the result is concatenated with the global context produced by the state-space fusion branch to form boundary-enhanced features for subsequent segmentation.

More »

Fig 2 Expand

Fig 3.

The structure of the Multi-scale Lesion Shape Prior module is illustrated, where the input features first pass through two stages of IDCNN to model long-range local context and are then transformed linearly to form the main decoding features, while the lower branch extracts shape-related responses through an edge-enhancement operator and convolutional networks to generate a global gating vector θ for channel importance modulation.

In the bottom pathway, the decoder features are processed by three scale-projection operators to obtain multi-scale shape representations, which are concatenated and fused through gating to produce shape-enhanced features, thereby injecting explicit contour and scale priors into the decoding stage.

More »

Fig 3 Expand

Fig 4.

Example image of ISIC dataset.

More »

Fig 4 Expand

Fig 5.

Example image of HAM10000 dataset.

More »

Fig 5 Expand

Fig 6.

Example image of PH2 dataset.

More »

Fig 6 Expand

Table 2.

Dataset statistics and train/test splits used in this study.

More »

Table 2 Expand

Table 3.

Hardware environment and training hyperparameters used in this study.

More »

Table 3 Expand

Table 4.

Comparison with state-of-the-art segmentation models on the ISIC 2018 dataset.

More »

Table 4 Expand

Table 5.

Comparison with state-of-the-art segmentation models on the HAM10000 dataset.

More »

Table 5 Expand

Table 6.

Comparison with state-of-the-art segmentation models on the PH2 dataset.

More »

Table 6 Expand

Fig 7.

Loss function changes with epoch.

More »

Fig 7 Expand

Table 7.

Systematic ablation study on ISIC 2018, HAM10000, and PH2 datasets.

BP denotes the boundary prior (gradient-driven boundary enhancement), HA denotes the hybrid attention in the encoder, and SP denotes the shape prior introduced by MLSP.

More »

Table 7 Expand

Fig 8.

Qualitative results of ablation experiments.

More »

Fig 8 Expand

Fig 9.

This figure illustrates the pairwise significance test results of the base model and different module combinations relative to the final model on three datasets.

By visualizing the p-value matrix, the statistical contribution of each module to the performance differences can be intuitively assessed, and the significance of the performance improvement can be verified.

More »

Fig 9 Expand

Table 8.

Hyperparameter sensitivity analysis of loss weights , , and .

More »

Table 8 Expand

Fig 10.

Qualitative Experimental Results of the ISIC2018 Dataset.

More »

Fig 10 Expand

Fig 11.

Qualitative Experimental Results of the HAM10000 Dataset.

More »

Fig 11 Expand

Fig 12.

Qualitative Experimental Results of the PH2 Dataset.

More »

Fig 12 Expand

Fig 13.

Experimental results of Grad-cam on three datasets.

More »

Fig 13 Expand

Fig 14.

The curves showing the changes in IoU, DSC, HD95, and ASD metrics of the proposed model on the ISIC 2018, HAM10000, and PH2 datasets under different Gaussian noise intensities demonstrate its robustness to lesion segmentation performance under random perturbations.

More »

Fig 14 Expand

Table 9.

Cross-domain generalization results on skin lesion segmentation.

More »

Table 9 Expand

Fig 15.

PR curve experimental results.

More »

Fig 15 Expand