Human Evaluation of Classification Errors on CUB-200 (test)
90.99Human Accuracy (%)top-1
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| top-1AI Correctness=AI is Correct2023.08 | 90.99 | 0.0773 | 465 | |
| PCNNAI Correctness=AI is Correct2023.08 | 79.78 | 0.1446 | 406 | |
| PCNNAI Correctness=Overall2023.08 | 64.58 | 0.0806 | 810 | |
| top-1AI Correctness=Overall2023.08 | 54.55 | 0.0954 | 990 | |
| PCNNAI Correctness=AI is Wrong2023.08 | 49.31 | 0.1727 | 404 | |
| top-1AI Correctness=AI is Wrong2023.08 | 22.28 | 0.1502 | 525 |