Scene Text Detection On Icdar 2013
평가 지표
F-Measure
Precision
Recall
평가 결과
이 벤치마크에서 각 모델의 성능 결과
비교 표
| 모델 이름 | F-Measure | Precision | Recall |
|---|---|---|---|
| scene-text-detection-with-supervised-pyramid | 92.1% | 93.8 | 90.5 |
| pixellink-detecting-scene-text-via-instance | 88.1% | 88.6 | 87.5 |
| synthetic-data-for-text-localisation-in | 83.0% | 92.0 | 75.5 |
| efficient-scene-text-localization-and | 77.1% | 81.8 | 72.4 |
| single-shot-text-detector-with-regional | 87% | 88 | 86 |
| unsharp-masking-layer-injecting-prior | 80.40% | - | - |
| multi-oriented-scene-text-detection-via | 88% | 92 | 84.4 |
| character-region-awareness-for-text-detection | - | 97.4 | 93.1 |
| detecting-oriented-text-in-natural-images-by | 85.3% | 87.7 | 83 |
| stn-ocr-a-single-neural-network-for-text | 90.3% | - | - |
| wordsup-exploiting-word-annotations-for | 90.34% | 93.34 | 87.53 |
| textboxes-a-single-shot-oriented-scene-text | 88%% | 91 | 84 |
| textfusenet-scene-text-detection-with-richer | 94.61% | 97.27 | 92.09 |
| reading-text-in-the-wild-with-convolutional | 76.8% | 88.5 | 67.8 |
| detecting-multi-oriented-text-with-corner | 87.6%% | 91.9 | 83.9 |
| mask-textspotter-an-end-to-end-trainable | 91.7% | 95 | 88.6 |