基于掩码伯努利扩散的皮肤病变分割模型
皮肤癌是一种常见且危及生命的疾病,早期检测能够显著提高存活率。为了提高诊断的精确性和效率,本文提出了一种基于掩码伯努利扩散的皮肤病变分割模型,在精确分割皮肤病灶方面具有优势。在ISIC2017和ISIC2018数据集上的实验结果验证了模型的有效性。具体来说,该模型在ISIC2017数据集上实现了81.42\%的平均交并比（mIoU）,而在ISIC2018数据集上达到了84.06\%的平均交并比（mIoU）。这些结果不仅展示了基于掩码伯努利扩散的模型在皮肤病变分割任务中的高精度,也说明了它在实际应用中的潜力,为皮肤癌的早期诊断和治疗提供了有力的技术支持。
Skin cancer is a common and life-threatening disease, and early detection can significantly increase survival rates. To enhance the accuracy and efficiency of diagnosis, this paper introduces a skin lesion segmentation model based on Bernoulli Diffusion of Mask, which excels in accurately segmenting skin lesions. Experimental results on the ISIC2017 and ISIC2018 datasets validate the effectiveness of the model. Specifically, the model achieved a mean Intersection over Union (mIoU) of 81.42\% on the ISIC2017 dataset and 84.06\% on the ISIC2018 dataset. These results not only demonstrate the high precision of the model in skin lesion segmentation tasks but also highlight its potential in practical applications, providing robust technical support for the early diagnosis and treatment of skin cancer.
徐士彪、郭莉、陈博裕
北京邮电大学人工智能学院，北京 100876北京邮电大学人工智能学院，北京 100876北京邮电大学人工智能学院，北京 100876
皮肤病学、性病学医学研究方法
生物医学工程皮肤病变分割扩散模型
Biomedical Engineering Skin Lesion Segmentation Diffusion Model
徐士彪,郭莉,陈博裕.基于掩码伯努利扩散的皮肤病变分割模型[EB/OL].(2025-04-16)[2025-07-16].http://www.paper.edu.cn/releasepaper/content/202504-149.点此复制
评论