论文泛读:《Automatic Skin Lesion Segmentation on Dermoscopic Images by the Means of Superpixel Merging》

简 介超像素合并算法superpixel merging)用于皮肤组织分割, 做到了 非深度方法 中的 SOTA,但是无法处理包含多个目标区域的情况,所以有很大改进空间(我已经想到了一个绝妙的办法,但是这里空间太...哇靠,还挺大的,好吧,我没想到),感兴趣的读者可以开展这个工作(论文代码已开源,数据公开但是图片很恶心)。值得一提的是,这个算法的效果跟 deep learning based SOTA 相差甚远。

Automatic Skin Lesion Segmentation on Dermoscopic Images by the Means of Superpixel Merging

论文泛读:《Automatic Skin Lesion Segmentation on Dermoscopic Images by the Means of Superpixel Merging》_第1张图片

论文泛读:《Automatic Skin Lesion Segmentation on Dermoscopic Images by the Means of Superpixel Merging》_第2张图片

论文泛读:《Automatic Skin Lesion Segmentation on Dermoscopic Images by the Means of Superpixel Merging》_第3张图片

论文泛读:《Automatic Skin Lesion Segmentation on Dermoscopic Images by the Means of Superpixel Merging》_第4张图片

论文泛读:《Automatic Skin Lesion Segmentation on Dermoscopic Images by the Means of Superpixel Merging》_第5张图片

论文泛读:《Automatic Skin Lesion Segmentation on Dermoscopic Images by the Means of Superpixel Merging》_第6张图片

你可能感兴趣的:(论文泛读:《Automatic Skin Lesion Segmentation on Dermoscopic Images by the Means of Superpixel Merging》)