Codesota · Benchmark · read2016(line-level)Home/Leaderboards/Vision & Documents/Document OCR/read2016(line-level)
Unknown

read2016(line-level).

read2016(line-level) is a state-of-the-art machine learning benchmark indexed on Codesota. This page tracks published model results, top scores per metric, and the SOTA timeline for read2016(line-level).

Paper ↗Leaderboard ↓
§ 01 · SOTA history

Year over year.

§ 02 · Leaderboard

Results by metric.

Found a wrong score or missing run?
Use row edits to send a sourced correction into moderation.
Add / edit result ↗Report issue ↗

Test Wer

Test Wer is the reported evaluation metric for read2016(line-level). Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for Test Werverifiedpapervendorcommunityunverified
RankModelTrustScoreYearLinksFix
01Span
From paper: SPAN: a Simple Predict & Align Network for Handwritten Paragraph Recognition
verified21.12021Paper ↗Code ↗Looks wrong?
02DAN
From paper: DAN: a Segmentation-free Document Attention Network for Handwritten Document Recognition
verified17.62022Paper ↗Code ↗Looks wrong?
03HTR-VT
From paper: HTR-VT: Handwritten Text Recognition with Vision Transformer
verified16.52024Paper ↗Code ↗Looks wrong?
04VAN
From paper: End-to-end Handwritten Paragraph Text Recognition Using a Vertical Attention Network
verified16.32020Paper ↗Code ↗Looks wrong?
05HTR-ConvText
Table 4: HTR-ConvText achieves WER 15.7% on READ2016 line-level test set. New SOTA, surpassing HTR-VT (16.5%) by 0.8 points.
verified15.72024Paper ↗Looks wrong?

Test Cer

Test Cer is the reported evaluation metric for read2016(line-level). Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for Test Cerverifiedpapervendorcommunityunverified
RankModelTrustScoreYearLinksFix
01CNN + BLSTM
From paper: Evaluating Sequence-to-Sequence Models for Handwritten Text Recognition
verified4.702019Paper ↗Code ↗Looks wrong?
02Span
From paper: SPAN: a Simple Predict & Align Network for Handwritten Paragraph Recognition
verified4.602021Paper ↗Code ↗Looks wrong?
03VAN
From paper: End-to-end Handwritten Paragraph Text Recognition Using a Vertical Attention Network
verified4.102020Paper ↗Code ↗Looks wrong?
04DAN
From paper: DAN: a Segmentation-free Document Attention Network for Handwritten Document Recognition
verified4.102022Paper ↗Code ↗Looks wrong?
05HTR-VT
From paper: HTR-VT: Handwritten Text Recognition with Vision Transformer
verified3.902024Paper ↗Code ↗Looks wrong?
06HTR-ConvText
Table 4: HTR-ConvText achieves CER 3.6% on READ2016 line-level test set (65.9M params). New SOTA, surpassing HTR-VT (3.9%) by 0.3 points.
verified3.602024Paper ↗Looks wrong?
§ 04 · Submit a result

Add to the leaderboard.

← Back to Document OCR