Table 1. Comparison of deep learning architectures for apoptosis classification.

Comparative table reporting accuracy, F1, and AUC metrics for a CNN, 3DCNN, Conv-LSTM, and Conv-Transformer. The classification accuracy is reported for static frames or image sequences. The last column shows which cell death study employed the same baseline architecture displayed in the table.

Classifier architecture	Frame accuracy	Sequence accuracy	F1	AUC	Study
CNN	74% ± 1.3	NA	0.77	0.779	La Greca et al., 2021; Verduijn et al., 2021
3DCNN	NA	91.22 % ± 0.15	0.91	0.924	-
Conv-LSTM	NA	97.42% ± 0.09	0.97	0.994	Kabir et al., 2022; Mobiny et al., 2020
Conv-Transformer	NA	98.27% ± 0.25	0.98	0.997	Our

CNN, convolutional neural network; NA, nonapplicable.