Skip to main content

Table 6 Results of SLI detection using state-of-the-art methods, classic deep neural networks and the proposed Sch-net

From: Sch-net: a deep learning architecture for automatic detection of schizophrenia

Method Accuracy Precision Recall F1-score
State-of-the-art method Grill [85,86,87] 0.9694 1.0000 0.9474 0.9730
Ramarao [88] 0.9941 - - -
Slogrove [89] 0.9800 0.9900 0.9900 0.9900
Sharma [90] 0.9790 - - -
Reddy [91] 0.9882 - - -
Deep Neural Network AlexNet 0.9132 0.9585 0.8810 0.9181
VGG16 0.9230 0.9897 0.8787 0.9309
ResNet34 0.9329 0.9489 0.9286 0.9386
DenseNet121 0.9461 0.9397 0.9643 0.9518
Xception 0.9622 0.9514 0.9863 0.9685
Sch-net (our) 0.9952 0.9979 0.9937 0.9958