Related Experiment Video
Updated: May 3, 2026
06:34
SCAnED - An Open-source Skin Segmentation Macro for Semi-automated Cell and Nuclei Detection in Epidermal and Dermal Skin Compartments
Published on: August 8, 2025
676
Multi-scale and edge-aware IFGNet for precise skin lesion segmentation in high-resolution dermoscopic images
Bo Li1, Peiwen Tan2, Jie Jia1
1The College of Electronic Information, Shanghai Dianji University, Shanghai, 201306, China.
Scientific Reports
|May 1, 2026
Summary
IFGNet, a novel hybrid deep learning model, enhances skin lesion segmentation accuracy for early melanoma diagnosis. This framework effectively addresses limitations in current methods by integrating CNNs and Transformers for improved diagnostic potential.
Area of Science:
- Medical image analysis
- Artificial intelligence in dermatology
Background:
- Accurate segmentation of dermoscopic images is crucial for early melanoma diagnosis.
- Current Convolutional Neural Network (CNN) and Transformer models have limitations in capturing both local details and global context, struggling with blurred boundaries and scale variations.
Purpose of the Study:
- To develop an advanced segmentation framework, IFGNet, that overcomes the limitations of existing methods for high-resolution skin lesion segmentation.
- To improve the accuracy and consistency of skin lesion segmentation for potential clinical application in computer-aided melanoma diagnosis.
Main Methods:
- IFGNet employs a hybrid CNN-Transformer architecture for synergistic feature extraction.
- The framework incorporates multi-scale convolution with large kernels and a boundary-aware decoding strategy.
- Key techniques include local-global feature fusion and a boundary refinement strategy to enhance lesion consistency.
Main Results:
- IFGNet demonstrated superior performance compared to state-of-the-art methods on the ISIC 2016, ISIC 2017, and ISIC 2018 benchmarks.
- The model achieved significant improvements in segmentation accuracy, measured by Dice and Intersection over Union (IoU) metrics.
- Results highlight IFGNet's effectiveness in handling blurred boundaries and scale variations in dermoscopic images.
Conclusions:
- IFGNet offers a robust solution for accurate high-resolution skin lesion segmentation.
- The proposed framework shows significant potential for advancing computer-aided diagnosis in clinical settings, particularly for early melanoma detection.
