Visual Question Answering on DAQUAR single-word answers portion
60.27AccuracyHUMAN
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| HUMAN2015.05 | 60.27 | 0.6104 | 0.7896 | |
| FULL2015.05 | 36.94 | 0.4815 | 0.8268 | |
| 2-VIS+BLSTM2015.05 | 35.78 | 0.4683 | 0.8215 | |
| ASK-NEURON2015.05 | 34.68 | 0.4076 | 0.7954 | |
| VIS+LSTM2015.05 | 34.41 | 0.4605 | 0.8223 | |
| IMG+BOW2015.05 | 34.17 | 0.4499 | 0.8148 | |
| LSTM2015.05 | 32.73 | 0.435 | 0.8162 | |
| BOW2015.05 | 32.67 | 0.4319 | 0.813 | |
| K-NNK=31, feature_dimension=132015.05 | 31.85 | 0.4242 | 0.8063 | |
| GUESS2015.05 | 18.24 | 0.2965 | 0.7759 | |
| MULTI-WORLD2015.05 | 12.73 | 0.181 | 0.5147 |