基于深度主动学习的甲状腺癌病理图像分类方法

张萌, 韩冰, 王哲, 尤富生, 李浩然

Papillary thyroid carcinoma pathological image classification based on deep active learning

Meng Zhang, Bing Han, Zhe Wang, Fusheng You, Haoran Li

表1 不同标注比例的分类结果（

λ_{1}

∶

λ_{2}

）

Table 1 Classification results with different ratio for

λ_{1} : λ_{2}

$λ_{1} : λ_{2}$	1%	5%	10%	20%	30%	40%	50%
0∶10	0.822	0.834	0.887	0.894	0.905	0.904	0.912
1∶9	0.862	0.879	0.893	0.898	0.914	0.918	0.919
2∶8	0.821	0.859	0.882	0.891	0.904	0.913	0.912
3∶7	0.830	0.849	0.856	0.859	0.899	0.914	0.914
4∶6	0.859	0.860	0.892	0.901	0.904	0.907	0.908
5∶5	0.831	0.840	0.868	0.879	0.888	0.905	0.907
6∶4	0.832	0.825	0.860	0.865	0.904	0.908	0.915
7∶3	0.833	0.844	0.843	0.899	0.890	0.898	0.911
8∶2	0.840	0.851	0.879	0.886	0.906	0.908	0.913
9∶1	0.835	0.840	0.863	0.893	0.900	0.912	0.916
10∶0	0.835	0.837	0.845	0.852	0.866	0.902	0.917