Table 2 Evaluation metrics obtained from the ablation experiments performed on the two datasets.

From: Tongue shape classification based on IF-RCNet

Dataset

Model

Mixed input

Feature fusion

Accuracy

F1 score

Our

RCA-Net

  

74.75%

73.61%

RCA-Net + Mixed input

 

75.50%

73.91%

RCANet + Feature fusion

 

75.75%

74.47%

RCA-Net + Mixed input + Feature fusion

76.87%

76.29%

BioHit

RCA-Net

  

72.19%

70.50%

RCA-Net + Mixed input

 

77.17%

78.92%

RCA-Net + Feature fusion

 

75.61%

72.76%

RCA-Net + Mixed input + Feature fusion

84.87%

84.22%