MSCRHO-Net: A Deep Learning Model for Effective Removal of Hair Occlusion from DermoscopicImages
Du Hongxuan1, Liu Qingyi1, Ren Yande2, Wang Yan1, Zhang Yanan1, Bai Peirui1#*
1(College of Electronic and Information Engineering, Shandong University of Science and Technology, Qingdao 266590, Shandong, China) 2(Department of Radiology, Affiliated Hospital of Qingdao University, Qingdao 265000,Shandong,China)
Abstract：It is of great significance to carry out early detection and diagnosis of skin cancer based on automatic analysis of dermoscopic images. However, hair occlusion poses a challenge to image feature extraction and skin lesion diagnosis. In this paper, a multi-scale cascade deep learning model (MSCRHO-Net) by integrating the Laplace pyramid was proposed. First, Laplace pyramid was employed to extract the key features of different scales in image space. A cascade block was designed for each scale channel to predict the hairless image by a coarse to fine scheme. High precision hair extraction and boundaries details retention were achieved through this operation. Then, a combined loss function including perceptual loss and SSIM loss was constructed, which was helpful to enhance details recovery and obtain more clear hair removal images. The performance of MSCRHO-Net was validated on synthetic dataset and real dataset ISIC2017(4750 training images, 400 synthetic test images, and 223 real test images). The experimental results demonstrated that MSCRHO-Net was able to remove hair effectively without learning of huge parameters.The mean values of SSIM and PSNR reached 0.958 4 and 35.49 respectively, which significantly improved the performance (P<0.05) compared with other traditional hair removal methods. MSCRHO-Net shows high adaptability and robustness to complex hair structure, and can deal with complicated scenarios such as damaged lesion texture and blurred image.
杜红萱, 刘庆一, 任延德, 王艳, 张亚楠, 白培瑞. MSCRHO-Net:一种有效去除皮肤镜图像毛发遮挡的深度学习模型[J]. 中国生物医学工程学报, 2024, 43(6): 662-672.
Du Hongxuan, Liu Qingyi, Ren Yande, Wang Yan, Zhang Yanan, Bai Peirui. MSCRHO-Net: A Deep Learning Model for Effective Removal of Hair Occlusion from DermoscopicImages. Chinese Journal of Biomedical Engineering, 2024, 43(6): 662-672.
[1] Li Wei, Raj ANJ, Tjahjadi T, et al. Digital hair removal by deep learning for skin lesion segmentation[J]. Pattern Recognition, 2021,117: 1-15. [2] Arnold M, Singh D, Laversanne M, et al. Global burden of cutaneous melanoma in 2020 and projections to 2040[J]. JAMA Dermatol, 2022, 158(5): 495-503. [3] 李航,余镇,倪东,等. 基于深度残差网络的皮肤镜图像黑色素瘤的识别[J]. 中国生物医学工程学报,2018, 37(3): 274-282. [4] 杨怡光,王钧程,谢凤英,等. 皮肤疾病辅助诊断系统中的数据与方法[J]. 协和医学杂志,2023, 14(1): 168-176. [5] Kaur R, Gholam HH, Sinha R. Hairlines removal and low contrast enhancement of melanoma skin images using convolutional neural network with aggregation of contextual information[J]. Biomed Signal Process and Control, 2022, 76(76): 1-11. [6] Lee T, Ng V, Gallagher R, et al. Dullrazor®: a software approach to hair removal from images[J]. Computers in Biology and Medicine, 1997, 27(6): 533-543. [7] Schmid-Saugeon P, Guillodb J, Thirana J. Towards a computer-aided diagnosis system for pigmented skin lesions[J]. Computerized Medical Imaging and Graphics, 2003, 27(1): 65-78. [8] Xie Fengying, Qin Shiyin, Jiang Zhiguo, et al. PDE-based unsupervised repair of hair-occluded information in dermoscopy images of melanoma[J]. Computerized Medical Imaging and Graphics, 2009, 33(4): 275-282. [9] Fiorese M, Peserico E, Silletti A, et al. VirtualShave: automated hair removal from digital dermatoscopic images[C]//Annual International Conference of the IEEE Engineering in Medicine and Biology Society. Boston: IEEE,2011: 5145-5148. [10] Delibasis K, Moutselos K, Vorgiazidou E, et al. Automated hair removal in dermoscopy images using shallow and deep learning neural architectures[J]. Computer Methods and Programs in Biomedicine Update, 2023,4(2666): 1-13. [11] Toossi MTB, Pourreza HR, Zare H, et al. An effective hair removal algorithm for dermoscopy images[J]. Skin Research and Technology, 2013, 19(3): 230-235. [12] Hasan SN, Gezer M, Azeez RA, et al. Skin lesion segmentation by using deep learning techniques[J]. 2019 Medical Technologies Congress (TIPTEKNO),2019, 11: 1-4. [13] Kiani K, Sharafat AR. E-shaver: An improved DullRazor® for digitally removing dark and light-colored hairs in dermoscopic images[J]. Computers in Biology and Medicine, 2011, 41(3): 139-145. [14] Xie Fengying, Li Yang, Meng Rusong, et al. No-reference hair occlusion assessment for dermoscopy images based on distribution feature[J]. Computers in Biology and Medicine, 2015, 59: 106-115. [15] Du Xian, Lee I, Anthony B. Hair segmentation using adaptive threshold from edge and branch length measures[J]. Computers in Biology and Medicine, 2017, 89: 314-324. [16] Koehoorn J, Sobiecki AC, Boda D, et al. Automated digital hair removal by threshold decomposition and morphological analysis[C]// International Symposium on Mathematical Morphology and its Applications to Signal and Image Processing. Reykjavik: Springer International Publishing, 2015: 15-26. [17] Nguyen NH, Lee TK, Atkins MS, et al. Segmentation of light and dark hair in dermoscopic images: a hybrid approach using a universal kernel[J]. Medical Imaging 2010: Image Processing, 2010, 7623(12): 1-8. [18] Abbas Q, Celebi ME, Garcia IF, et al. Hair removal methods: a comparative study for dermoscopy images[J]. Biomedical Signal Processing and Control, 2011, 6(4): 395-404. [19] Huang A, Kwan SY, Chang WY, et al. A robust hair segmentation and removal approach for clinical images of skin lesions[C]// Annual International Conference of the IEEE Engineering in Medicine and Biology Society. Osaka: IEEE, 2013: 3315-3318. [20] 胡明娣,吴怡,宋尧,等. 单幅图像去雨数据集和深度学习算法的联合评估与展望[J]. 中国图象图形学报,2022, 27(5), 1359-1391. [21] 赵蓉,陈平,潘晋孝. 基于多尺度扩张残差循环网络的图像去雨算法[J]. 中国科技论文,2023, 18(3): 310-316. [22] 张世辉,路佳琪,宋丹丹,等. 基于多尺度特征结合细节恢复的单幅图像去雾方法[J]. 电子与信息学报,2022, 44(11): 3967-3976. [23] 李炜. 基于皮肤镜图像的数字毛发去除技术研究[D].汕头: 汕头大学, 2023. [24] Attia M, Hossny M, Zhou Hailing, et al. Digital hair segmentation using hybrid convolutional and recurrent neural networks architecture[J]. Computer Methods and Programs in Biomedicine, 2019, 177: 17-30. [25] Talavera ML, Bibiloni P, Gonzalez HM. Hair segmentation and removal in dermoscopic images using deep learning[J]. IEEE ACCESS, 2021, 9: 2694-2704. [26] Lai Weisheng., Huang Jiabin, Ahuja N, et al. Fast and accurate image super-resolution with deep Laplacian pyramid networks[J]. IEEE Transactions on Pattern Analysis And Machine Intelligence, 2019, 41(11): 2599-2613. [27] Cheng Hongquan, Wu Huayi, Zheng Jie, et al. A hierarchical self-attention augmented Laplacian pyramid expanding network for change detection in high-resolution remote sensing images[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 182: 52-66. [28] Tai Ying, Yang Jian, Liu Xiaoming. Image super-resolution via deep recursive residual network[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 3147-3155. [29] Fu Xueyang, Liang Borong, Huang Yue, et al. Lightweight pyramid networks for image deraining[J]. IEEE Transactions on Neural Networks and Learning Systems, 2020, 31(6): 1794-1807. [30] Wang Zhou, Bovik AC, Sheikh HR, et al. Image quality assessment: from error visibility to structural similarity[J]. IEEE Transactions on Image Processing, 2004, 13(4): 600-612. [31] Johnson J, Alahi A, Li Feifei. Perceptual losses for real-time style transfer and super-resolution[J].CoRR, 2016,1063(8155): 694-711. [32] 龙霄潇,程新景,朱昊,等. 三维视觉前沿进展[J]. 中国图象图形学报,2021, 26(6): 1389-1428. [33] Attia M, Hossny M, Zhou H, et al. Realistic hair simulator for skin lesion images: a novel benchemarking tool[J]. Artificial Intelligence In Medicine, 2020, 108: 1-12. [34] Fu Yan, Wang Shengchun. A no reference image quality assessment metric based on visual perception[J]. Algorithms, 2016,9(4): 87-97. [35] Pathan S, Prabhu KG, Siddalingaswamy PC. Techniques and algorithms for computer aided diagnosis of pigmented skin lesions—a review[J]. Biomedical Signal Processing and Control, 2018, 39: 237-262.
