This section lists current results on the CoverDB benchmark. All results must be computed using this test data.
To add your results, send the following information to xbaro@cvc.uab.es
· Full citation for your paper.
· Optionally: bibtex entry, pdf of paper (or link), webpage link.
· TASK 1 Performance Value: (Detected text area ∩ ground truth text area) / (Detected text area ∪ ground truth text area), computed over the provided test set.
· TASK 2 Performance Curve: Performance value as a function of dataset size.
· TASK 3 Performance Curve: Performance value as a function of dataset size.
Results submitted to this benchmark
Authors | Performance | Paper | Code |
S. Escalera, X. Baró, J. Vitrià, and P. Radeva | 28 ± 2.36 % | link | link |
J. C. Pérez and A. J. Pérez | 16 ± 13.9 % | link | N/A |