spin

SPIN

1 Introduction

This code repository contains an implementation of (SPIN: SPIN: Structure-Preserving Inner Offset Network for Scene Text Recognition (AAAI2021) ). Chromatic diffculties in complex scenes have not been paid much attention on. We introduce a new learnable geometric-unrelated module,the Structure-Preserving Inner Offset Network (SPIN), which allows the color manipulation of source data within the network.

2. Preparing Dataset

Train Dataset

Dataset	Samples	Description	Release
MJSynth	8919257	Scene text recognition synthetic data set	Link
SynText	7266164	A synthesized by scene text dataset, and the text is cropped from the large image	Link

Validation Dataset

Test Set	Instance Number	Note
IIIT5K	3000	regular
SVT	647	regular
IC03_860	860	regular
IC13_857	857	regular
IC15_1811	1811	irregular
SVTP	645	irregular
CUTE80	288	irregular

Test Dataset

Test Set	Instance Number	Note
IIIT5K	3000	regular
SVT	647	regular
IC03_860	860	regular
IC13_857	857	regular
IC15_1811	1811	irregular
SVTP	645	irregular
CUTE80	288	irregular

3 Getting Started

Preparation

A quick start is to use above lmdb-formatted datasets that contain the full benchmarks for scene text recognition tasks as belows.

Data Type: LMDB

File storage format:
   |-- train           
   |   |-- MJ
   |   |-- ST
   |-- validation
   |   |-- mixture
   |-- evaluation
   |   |-- mixture

Training

Run the following bash command in the command line,

cd .
bash ./train.sh 

We provide the implementation of online validation. If you want to close it to save training time, you may modify the startup script to add --no-validate command.

Evaluation

Run following scripts to compare different rectification modules.

cd .
bash ./test_scripts/test_affine.sh

cd .
bash ./test_scripts/test_tps.sh

cd .
bash ./test_scripts/test_spin.sh

cd .
bash ./test_scripts/test_gaspin.sh

4 Results

Evaluation

Methods	Regular Text				Irregular Text			Download
Name	IIIT5K	SVT	IC03	IC13	IC15	SVTP	CUTE80	Config	Model
Affine(Report)	-	-	-	-	-	-	-	-	-
Affine	94.0	87.9	93.6	94.4	81.2	80.9	83.0	Config	pth [Link] (Access Code: 5nr1)
TPS(Report)	87.9	87.5	94.9	93.6	77.6	79.2	74.0	-	-
TPS	94.2	90.4	94.5	95.0	82.1	82.6	83.7	Config	pth [Link] (Acceess Code: 024F)
SPIN(Report)	94.7	87.6	93.4	91.5	79.1	79.7	85.1	-	-
SPIN	94.7	89.8	93.4	94.1	80.7	82.2	83.7	Config	pth [Link] (Code:M45z)
GA-SPIN(Report)	94.7	90.3	94.4	92.8	82.2	82.8	87.5	-	-
GA-SPIN	94.6	89.0	93.3	94.2	80.7	83.0	84.7	Config	pth [Link] (Access Code: 12q3)

Visualization

Here is the picture for result visualization.

Citation

@article{SPIN,
  title={SPIN: Structure-Preserving Inner Offset Network for Scene Text Recognition},
  author={Chengwei Zhang, Yunlu Xu, Zhanzhan Cheng, Shiliang Pu, Yi Niu, Fei Wu},
  journal={AAAI},
  year={2021}
}

License

This project is released under the Apache 2.0 license

Contact

If there is any suggestion and problem, please feel free to contact the author with [email protected].