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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The subtleties of human perception, as measured by vision scientists throughthe use of psychophysics, are important clues to the internal workings ofvisual recognition. For instance, measured reaction time can indicate whether avisual stimulus is easy for a subject to recognize, or whether it is hard. Inthis paper, we consider how to incorporate psychophysical measurements ofvisual perception into the loss function of a deep neural network being trainedfor a recognition task, under the assumption that such information can enforceconsistency with human behavior. As a case study to assess the viability ofthis approach, we look at the problem of handwritten document transcription.While good progress has been made towards automatically transcribing modernhandwriting, significant challenges remain in transcribing historicaldocuments. Here we describe a general enhancement strategy, underpinned by thenew loss formulation, which can be applied to the training regime of any deeplearning-based document transcription system. Through experimentation, reliableperformance improvement is demonstrated for the standard IAM and RIMES datasetsfor three different network architectures. Further, we go on to showfeasibility for our approach on a new dataset of digitized Latin manuscripts,originally produced by scribes in the Cloister of St. Gall in the the 9thcentury.

Measuring Human Perception to Improve Handwritten Document Transcription