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 is a major computer vision task with a big set of associatedchallenges. One of those traditional challenges is the coupled nature of textrecognition and segmentation. This problem has been progressively solved overthe past decades, going from segmentation based recognition to segmentationfree approaches, which proved more accurate and much cheaper to annotate datafor. We take a step from segmentation-free single line recognition towardssegmentation-free multi-line / full page recognition. We propose a novel andsimple neural network module, termed \textbf{OrigamiNet}, that can augment anyCTC-trained, fully convolutional single line text recognizer, to convert itinto a multi-line version by providing the model with enough spatial capacityto be able to properly collapse a 2D input signal into 1D without losinginformation. Such modified networks can be trained using exactly their samesimple original procedure, and using only \textbf{unsegmented} image and textpairs. We carry out a set of interpretability experiments that show that ourtrained models learn an accurate implicit line segmentation. We achievestate-of-the-art character error rate on both IAM \& ICDAR 2017 HTR benchmarksfor handwriting recognition, surpassing all other methods in the literature. OnIAM we even surpass single line methods that use accurate localizationinformation during training. Our code is available online at\url{https://github.com/IntuitionMachines/OrigamiNet}.

OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold