Research Article
Conceptual Cognitive Modeling for Fine-Grained Annotation Quality Assessment of Object Detection Datasets
Table 6
Annotation accuracy of human annotations for the selected images of the PASCAL 2007 dataset (the average values are computed for the twenty classes).
| Class | Flat accuracy of label | Hierarchical accuracy of label | Precision | Recall | Precision | Recall |
| Person | 0.98 ± 0.01 | 0.92 ± 0.05 | 0.99 ± 0.01 | 0.92 ± 0.04 | Car | 0.99 ± 0.02 | 0.94 ± 0.04 | 0.99 ± 0.01 | 0.94 ± 0.03 | Chair | 0.96 ± 0.03 | 0.74 ± 0.08 | 0.98 ± 0.01 | 0.82 ± 0.07 | Bottle | 0.98 ± 0.01 | 0.81 ± 0.08 | 0.99 ± 0.01 | 0.82 ± 0.07 | Potted plant | 1 ± 0 | 0.54 ± 0.31 | 1 ± 0 | 0.57 ± 0.29 | Cow | 0.99 ± 0.01 | 0.95 ± 0.01 | 0.997 ± 0.005 | 0.96 ± 0.01 | Dining table | 0.75 ± 0.16 | 0.59 ± 0.18 | 0.91 ± 0.06 | 0.64 ± 0.15 | Bus | 1 ± 0 | 0.94 ± 0.04 | 1 ± 0 | 0.96 ± 0.03 | On average | 0.96 ± 0.01 | 0.89 ± 0.04 | 0.983 ± 0.005 | 0.90 ± 0.04 |
|
|