Table 1 Metrics of different ANN methods (MIL with attention (MIL-attention), MIL with maxpooling (MIL-max), MIL with mean pooling (MIL-mean), and the baseline SELU CNN (baseline SELU)).

From: Artificial neural networks and pathologists recognize basal cell carcinomas based on different histological patterns

Data type

Method

Accuracy

F1-score

AUC

Sensitivity

Specificity

tiles

MIL-attention

0.95 (0.94–0.96)

0.97 (0.96–0.97)

0.99 (0.99–0.99)

0.96 (0.95–0.97)

0.93 (0.90–0.95)

tiles

MIL-max

0.90 (0.89–0.92)

0.93 (0.92–0.94)

0.96 (0.96–0.96)

0.94 (0.91–0.98)

0.78 (0.69–0.87)

tiles

MIL-mean

0.88 (0.87–0.89)

0.92 (0.91–0.92)

0.91 (0.91–0.92)

0.93 (0.92–0.95)

0.72 (0.69–0.75)

downscaled WSIs

baseline SELU

0.76 (0.72–0.80)

0.83 (0.78–0.87)

0.84 (0.82–0.86)

0.78 (0.68–0.87)

0.73 (0.57–0.89)

  1. AUC area under the ROC curve, CNN convolutional neural network, MIL multiple instance learning, WSI whole-slide image.
  2. Bold values indicate name and metrics of the proposed attention-ANN method.
Back to article page