method: SAR2018-12-19
Authors: Hui Li, Peng Wang, Chunhua Shen, Guyu Zhang
Description: We propose an easy-to-implement strong baseline for irregular scene text recognition, using off- the-shelf neural network components and only word-level annotations. It is composed of a 31-layer ResNet, an LSTM- based encoder-decoder framework and a 2-dimensional attention module. Despite its simplicity, the proposed method is robust. It achieves state-of-the-art performance on irregular text recognition benchmarks and comparable results on regular text datasets.
method: SAR_tensorflow_reproduced2020-01-11
Authors: Pay20Y
Description: SAR reimplemented in tensorflow and trained with Synth90K and SynthText
method: SAR tf-reproduce2019-10-14
Authors: Kolorin
Description: The tensorflow reproduce of SAR(https://www.aaai.org/ojs/index.php/AAAI/article/view/4881)
Date | Method | Total Edit distance (case sensitive) | Correctly Recognised Words (case sensitive) | T.E.D. (case insensitive) | C.R.W. (case insensitive) | |||
---|---|---|---|---|---|---|---|---|
2018-12-19 | SAR | 437.1642 | 67.40% | 203.1446 | 78.82% | |||
2020-01-11 | SAR_tensorflow_reproduced | 1,031.1057 | 41.36% | 333.1592 | 69.19% | |||
2019-10-14 | SAR tf-reproduce | 843.4374 | 40.73% | 494.6210 | 55.66% | |||
2019-10-14 | transformer-based method | 807.5391 | 39.24% | 454.9708 | 54.45% |