Yiyao Wang
Text Recognition: Add script to evaluate text recognition by ICDAR2003 (#71)
46b1f95
|
raw
history blame
2.92 kB

CRNN

An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition

Results of accuracy evaluation with tools/eval at different text recognition datasets.

Model name ICDAR03(%) IIIT5k(%) CUTE80(%)
CRNN_EN 81.66 74.33 52.78
CRNN_EN_FP16 82.01 74.93 52.34
CRNN_CH 71.28 80.90 67.36
CRNN_CH_FP16 78.63 80.93 67.01

*: 'FP16' stands for 'model quantized into FP16'.

Note:

Demo

NOTE:

  • This demo uses text_detection_db as text detector.
  • Selected model must match with the charset:
    • Try text_recognition_CRNN_EN_2021sep.onnx with charset_36_EN.txt.
    • Try text_recognition_CRNN_CH_2021sep.onnx with charset_94_CH.txt
    • Try text_recognition_CRNN_CN_2021sep.onnx with charset_3944_CN.txt.

Run the demo detecting English:

# detect on camera input
python demo.py
# detect on an image
python demo.py --input /path/to/image

Run the demo detecting Chinese:

# detect on camera input
python demo.py --model text_recognition_CRNN_CN_2021nov.onnx --charset charset_3944_CN.txt
# detect on an image
python demo.py --input /path/to/image --model text_recognition_CRNN_CN_2021nov.onnx --charset charset_3944_CN.txt

Examples

CRNNCTC

demo

License

All files in this directory are licensed under Apache 2.0 License.

Reference