Image Classification on FER2013 Happy, Surprised and Neutral (test)
76.8Accuracy (Full)BCNN-ISLaB-FLOW
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| BCNN-ISLaB-FLOWInitial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 76.8 | 75.1 | — | — | — | |
| BNN-HORSEInitial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 68.4 | 68.4 | 3,370.5 | 3 | 3 | |
| Laplace-SpaMInitial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 65.8 | 64.7 | 5,608 | 3 | 3 | |
| BNN-CONCRETEInitial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 64.6 | — | 10 | 3 | 3 | |
| IS-ANN-L1Initial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 62.3 | 63.9 | 746.5 | 1 | 1 | |
| BCNN-ISLaB-LRTInitial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 61 | 56 | — | — | — | |
| BLR-FLOWInitial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 60.7 | 60.9 | 28.5 | 1 | 1 | |
| BLR-LRTInitial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 60.3 | 60.4 | 34 | 1 | 1 | |
| ISLaB-LRTInitial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 59.6 | 59.5 | 24 | 1 | 1 | |
| ISLaB-FLOWInitial weights=969’112, Hidden layers=2, Nodes per layer=2002025.03 | 58.4 | 58.3 | 24 | 1 | 1 |