ICDAR 2019 Chinese Signboard Text Recognition Challenge (Meituan Dataset)

Meituan’s new 25,000‑image Chinese signboard dataset—captured across varied devices, locations, lighting and angles—serves as the benchmark for the ICDAR 2019 Robust Reading Challenge, which features four tasks (end‑to‑end recognition, text‑line localization, single‑character and string recognition) with a 20k/2k/3k train/validation/test split, character‑level annotations, and participation from academic and industry partners, with registration opening March 1 and the competition concluding at the September conference in Sydney.

Meituan Technology Team
Meituan Technology Team
Meituan Technology Team
ICDAR 2019 Chinese Signboard Text Recognition Challenge (Meituan Dataset)

Meituan released a 25,000‑image dataset of real‑world Chinese signboards captured nationwide, providing a diverse benchmark for OCR research.

The images vary in device, location, lighting, and angle, making them ideal for evaluating algorithm robustness.

At ICDAR 2019 (Sydney, Sep 20‑25) the "Robust Reading Challenge on Reading Chinese Text on Signboards" will use this dataset, defining four tasks: end‑to‑end recognition, text‑line localization, single‑character recognition, and string recognition.

Data split: 20,000 training, 2,000 validation, 3,000 test images. Annotations include character‑level bounding boxes and transcriptions; Chinese dominates with 10‑30% English/digits.

Organizers: Meituan researchers (Wang Dong, Zhang Rui, Liu Xi, Zhou Yongsheng) and academic partners from Huazhong University of Science and Technology, Microsoft, Nanyang Technological University, University of Alicante, and IIIT Hyderabad.

Key dates: registration opens Mar 1, training set Mar 18, test set Apr 15, submission opens Apr 16, deadline Apr 30, final report May 10, conference Sep 20. Register at http://rrc.cvc.uab.es/?ch=12 or scan the QR code. Queries: [email protected] (subject "ICDAR2019").

Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

OCRChinese textICDAR2019
Meituan Technology Team
Written by

Meituan Technology Team

Over 10,000 engineers powering China’s leading lifestyle services e‑commerce platform. Supporting hundreds of millions of consumers, millions of merchants across 2,000+ industries. This is the public channel for the tech teams behind Meituan, Dianping, Meituan Waimai, Meituan Select, and related services.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.