基于小波池化Gabor卷积神经网络的皮肤癌分类
刘瑞欣 刘海华通讯作者

中南民族大学 生物医学工程学院，武汉，430074；

摘要：皮肤癌是目前人类最频发的癌症之一，严重影响人类的生命安全，为了降低皮肤癌的影响，对皮肤病的早期精确诊断和分类具有非常重要的意义。迄今为止，深度学习已被广泛应用于疾病的临床辅助诊断中，如卷积神经网络（CNN）在基于医学图像分类的疾病辅助诊断中表现出优越的性能。针对皮肤疾病图像的特点，本文采用Gabor滤波器对图像中纹理信息，边缘信息的敏感性，以及离散小波变换对噪声的鲁棒性，构建由Gabor卷积核和小波下采样层组成的Gabor卷积网络模型（WAPL-Gabor CNN）。该网络利用Gabor的实部和虚部建立卷积层，而利用注意力小波下采样构建下采样层。为了检验Gabor卷积网络的性能，采用HAM10000和Small-ISIC2018等公开数据集进行实验，其结果表明，与其它方法相比该网络模型非常有效，其分类准确率分别达到99.76%和91.6%。

关键词：皮肤癌；Gabor滤波器；注意力机制；离散小波变换

参考文献

[1] Deng X, Yin Q, Guo P. Efficient structural pseudoinverse learning-based hierarchical representation learning for skin lesion classification[J]. Complex & Intelligent Systems, 2022, 8(2): 1445-1457.

[2] Wu Q, Yu Y, Zhang X. A skin cancer classification method based on discrete wavelet down-sampling feature reconstruction[J]. Electronics, 2023, 12(9): 2103.

[3] Ji C, Su J. Gabor binary layer in convolutional neural networks[C]//2018 25th IEEE ICIP. 2018: 3408-3412.

[4] Liu C, Li J, He L, et al. Naive Gabor networks for hyperspectral image classification[J]. IEEE Transactions on Neural Networks and Learning Systems, 2020, 32(1): 376-390.

[5] Li Q, Shen L, Guo S, et al. WaveCNet: Wavelet Integrated CNNs to Suppress Aliasing Effect for Noise-Robust Image Classification[J]. IEEE Transactions on Image Processing, 2021, 30:7074-7089.

[6] Li S, Florencio D, Li W, et al. A Fusion Framework for Camouflaged Moving Foreground Detection in the Wavelet Domain[J]. IEEE Trans. Image Process., 2018, 27(8):3918-3930

[7] Tschandl P, Rosendahl C, Kittler H. The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions[J]. Scientific data, 2018, 5(1): 1-9

[8] Codella N, Rotemberg V, Tschandl P, et al. Skin lesion analysis toward melanoma detection 2018: A challenge hosted by the international skin imaging collaboration (isic)[J]. arXiv preprint arXiv:1902.03368, 2019.

[9] Lan Z, Cai S, He X, et al. FixCaps: An improved capsules network for diagnosis of skin cancer[J]. IEEE Access, 2022, 10: 76261-76267.

[10] Chanda D, Onim M S H, Nyeem H, et al. DCENSnet: A new deep convolutional ensemble network for skin cancer classification[J]. Biomedical Signal Processing and Control, 2024, 89: 105757.

[11] Sulthana R, Chamola V, Hussain Z, et al. A novel end to end deep convolutional neural network based skin lesion classification framework[J]. Expert Systems with Applications, 2024, 246: 123056.

[12] İlker G, Özkan İ. SADASNet: A Selective and Adaptive Deep Architecture Search Network with Hyperparameter Optimization for Robust Skin Cancer Classification[J]. Diagnostics, 2025, 15(5): 541.

[13] Dosovitskiy A, Beyer L, Kolesnikov A, et al. An image is worth 16x16 words: Transformers for image recognition at scale[J]. arXiv preprint arXiv:2010.11929, 2020.

[14] Liu S, Wang L, Yue W. An efficient medical image classification network based on multi branch CNN, token grouping Transformer and mixer MLP[J]. Applied Soft Computing, 2024, 153: 111323.

[15] Ren Z, Liu S, Wang L, et al. Conv SdMLPMixer: A hybrid medical image classification network based on multi branch CNN and multi scale multi dimensional MLP[J]. Information Fusion, 2025: 102937.