English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Text recognition remains a fundamental and extensively researched topic incomputer vision, largely owing to its wide array of commercial applications.The challenging nature of the very problem however dictated a fragmentation ofresearch efforts: Scene Text Recognition (STR) that deals with text in everydayscenes, and Handwriting Text Recognition (HTR) that tackles hand-written text.In this paper, for the first time, we argue for their unification -- we aim fora single model that can compete favourably with two separate state-of-the-artSTR and HTR models. We first show that cross-utilisation of STR and HTR modelstrigger significant performance drops due to differences in their inherentchallenges. We then tackle their union by introducing a knowledge distillation(KD) based framework. This is however non-trivial, largely due to thevariable-length and sequential nature of text sequences, which rendersoff-the-shelf KD techniques that mostly works with global fixed-length datainadequate. For that, we propose three distillation losses all of which arespecifically designed to cope with the aforementioned unique characteristics oftext recognition. Empirical evidence suggests that our proposed unified modelperforms on par with individual models, even surpassing them in certain cases.Ablative studies demonstrate that naive baselines such as a two-stageframework, and domain adaption/generalisation alternatives do not work as well,further verifying the appropriateness of our design.

Text is Text, No Matter What: Unifying Text Recognition using Knowledge Distillation