Medical Intelligence and Language Engineering Lab    TTS Demo   |   Downloads   |   Videos   |   Contact Us   
     Home   |    MRRC Home   |    Description   |    Dataset   |    Results   |    Publications      
 

Multi-script Robust Reading Competition (MRRC)

In the Indian scenario, we find multi-script hoardings and street name boards in multiple languages. The main theme of this competition will be to recognize the multi-script text from scenic images obtained from Indian roads. This will help in building novel applications for identification and recognition of Indic scripts in camera-captured scenic images. The competition will provide a platform for researchers around the globe to address this issue.

Description

A competition to detect mixed text in Kannada and/or English from scene images was organized by Medical Intelligence and Language Engineering (MILE) Lab in ICDAR 2013. The motivation was to look for script-independent algorithms that detect the text and extract it from the scene images, which may be applied directly to an unknown script.

We organized the competition for four distinct tasks:

  • Text localization from scene images
  • Text segmentation from scene images
  • English word recognition from cropped word images
  • Kannada word recognition from cropped word images

Dataset

By downloading and using the dataset, you agree to acknowledge their source and cite the paper given below in related publications.

Training Dataset

Test Dataset

Results

Text localization

ParticipantPrecisionRecallF-scoreAS (normal)AS (complex)
Yin et. al0.640.420.510.580.56

Text Segmentation

ParticipantPrecisionRecallF-scoreAS (normal)AS (complex)
Yin et. al0.710.670.690.800.56
Gomez et. al0.640.580.610.690.35
Sethi et. al0.330.720.450.650.18
OTCYMIST0.500.290.370.460.21

Word Recognition

MethodEWRTED-EKWRTED-K
Benchmark57.7215.111.1178.7
PLT46.9299.55.3210.6
MAPS46.9305.94.9209.1
NESP45.1305.45.8212.3
Baseline (raw image)37.5369.42.5218.5

For more details, please refer the publications.

Publications

Deepak Kumar, M. N. Anil Prasad and A. G. Ramakrishnan, “Multi-script robust reading competition in ICDAR 2013,” ACM - Proc. International Workshop on Multilingual OCR, (MOCR 2013), 24 Aug. 2013, Washington DC, USA.(Download)
© 2013 Medical Intelligence and Language Engineering Lab, EE, IISc, Bangalore.