ace

ACE Loss

1. Introduction

Aggregation cross-entropy (ACE: Aggregation Cross-Entropy for Sequence Recognition (CVPR 2019)), for sequence recognition from a brand new perspective. ACE loss only requires only characters and their numbers in the sequence annotation for supervision, which allows it to advance beyond sequence recognition

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

cd ../test_scripts
bash ./test_ace.sh

4. Results

Evaluation

Methods	Regular Text				Irregular Text			Download
Name	IIIT5K	SVT	IC03	IC13	IC15	SVTP	CUTE80	Config	Model
ACE Loss(Report)	82.3	82.6	92.1	-	-	-	-	-	-
ACE Loss	90.9	84.2	90.2	90.1	73.4	71.9	77.1	Config	pth [Link] (Access Code: 0i7j)

Visualization

Here is the picture for result visualization.

Citation

@inproceedings{ACE,
  author={Zecheng Xie and Yaoxiong Huang and Yuanzhi Zhu and Lianwen Jin and Yuliang Liu and Lele Xie},
  title={Aggregation Cross-Entropy for Sequence Recognition},
  booktitle={CVPR2019},
  pages={6538--6547},
  publisher={Computer Vision Foundation / {IEEE}},
  year={2019},
}

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 qiaoliang6@hikvision.com.