New Challenges for 2019 Announced
Six new challenges will be introduced for the 2019 edition of the Robust Reading Competition:
The ICDAR 2019 Robust Reading Challenge on Scene Text Visual Question Answering introduces a special Visual Question Answering challenge, with questions that require the correct extraction and understanding of scene text in order to be properly answered.
The ICDAR 2019 Robust Reading Challenge on Multi-lingual scene text detection and recognition is the second edition of the MLT Challenge and will introduce two new scripts/languages, new real and synthetic data and a complete end-to-end task that aims to the combined localisation, script identification and text recognition.
The ICDAR 2019 Robust Reading Challenge on Large-scale Street View Text with Partial Labeling is a large scale bi-lingual (Latin and Chinese) scene text extraction challenge, based on a large dataset of 25,000 fully annotated and 400,000 weakly annotated images.
The ICDAR 2019 Robust Reading Challenge on Arbitrary-Shaped Text is pushing the limits of text localisation by focusing on text laid out in challenging ways, including peculiar text envelopes and curved text lines. Contains Chinese and English text.
The ICDAR 2019 Robust Reading Challenge on Scanned Receipts OCR and Information Extraction revives a decades old favourite topic of administrative document image analysis: the extraction of information from scanned receipts.
Finally, the ICDAR 2019 Robust Reading Challenge on Reading Chinese Text on Signboard focuses on Chinese text found on signboards.