Semi-supervised tongue image segmentation method for traditional chinese medicine based on mutual learning with dual models

Table 1. Comparison of model performance（mIoU）under different strategies
View table
View in Article
Table 1. Comparison of model performance（mIoU）under different strategies
是否使用双模型是否使用互学习损失函数 1/100 1/50 1/25 1/8
× × 93.21 94.52 95.13 96.19
√ × 96.13 97.29 97.99 98.35
× √ 96.33 97.78 97.83 98.17
√ √ 96.67 97.92 98.52 98.85

Table 2. Comparison of mIoUs between the proposed method and other methods under different proportion of annotated data

View table

View in Article

Table 2. Comparison of mIoUs between the proposed method and other methods under different proportion of annotated data

方法	1/100	1/50	1/25	1/8
DeepLabv3+^［13］	85.53	87.02	91.06	94.88
ClassMix^［17］	94.74	96.14	97.35	98.09
Mean Teacher^［26］	87.12	89.52	92.28	95.61
CCT^［18］	83.71	87.98	92.02	95.37
S4GAN^［15］	90.31	94.74	96.69	97.61
ST++^［20］	94.62	95.09	96.08	97.31
本文方法	96.67	97.92	98.52	98.85

Tools

Get Citation

Copy Citation Text

Fangxu LI, Wangming XU, Xue XU, Yun JIA. Semi-supervised tongue image segmentation method for traditional chinese medicine based on mutual learning with dual models[J]. Chinese Journal of Liquid Crystals and Displays, 2024, 39(8): 1014

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Image Segmentation

Received: Sep. 22, 2023

Accepted: --

Published Online: Sep. 27, 2024

The Author Email: Yun JIA (jiayun@cug.edu.cn)

DOI:10.37188/CJLCD.2023-0308

Topics

Table 0. [in Chinese]

Table 0. [in Chinese]

Table 1. Comparison of model performance（mIoU）under different strategies

Table 1. Comparison of model performance（mIoU）under different strategies

Table 2. Comparison of mIoUs between the proposed method and other methods under different proportion of annotated data

Table 2. Comparison of mIoUs between the proposed method and other methods under different proportion of annotated data