Table 2. Stanford decoding accuracies (CMC AUC % (Cumulative Match Curve Area Under Curve)).
Feature Set | S1 | S2 | S3 | S4 | S5 | S6 | S7 | S8 | S9 | S10 | GS | Mean |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Visual | 59.27 | 56.91 | 57.28 | 57.51 | 62.57 | 60.96 | 56.59 | 54.5 | 54.11 | 61.08 | 61.97 | 58.08 |
Semantic | 60.12 | 56.89 | 61.38 | 60.35 | 65.49 | 63.93 | 62.36 | 55.93 | 58.07 | 64.28 | 64.82 | 60.88 |
Visuo-semantic | 61.86 | 58.45 | 60.8 | 61.85 | 67.9 | 66.08 | 63.07 | 56.59 | 57.84 | 65.77 | 66.01 | 62.02 |