Optical Character Recognition2020en

lam(line-level)

Dataset from Papers With Code

Metrics:accuracy, cer, wer, f1

test-cer

#	Model	Score	Paper / Code	Date
1	HTR-VT	2.8	HTR-VT: Handwritten Text Recognition with Vision Transformer Code	Sep 2024
2	OrigamiNet-24	3	OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold Code	Jun 2020
3	OrigamiNet-12	3.1	OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold Code	Jun 2020
4	OrigamiNet-18	3.1	OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold Code	Jun 2020
5	TrOCR	3.6	TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models Code	Sep 2021
6	GFCN	5.2	Recurrence-free unconstrained handwritten text recognition using gated fully convolutional network Code	Dec 2020

test-wer

#	Model	Score	Paper / Code	Date
1	HTR-VT	7.4	HTR-VT: Handwritten Text Recognition with Vision Transformer Code	Sep 2024
2	OrigamiNet-24	11	OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold Code	Jun 2020
3	OrigamiNet-18	11.1	OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold Code	Jun 2020
4	OrigamiNet-12	11.2	OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold Code	Jun 2020
5	TrOCR	11.6	TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models Code	Sep 2021
6	GFCN	18.5	Recurrence-free unconstrained handwritten text recognition using gated fully convolutional network Code	Dec 2020

Related Papers4

HTR-VT: Handwritten Text Recognition with Vision Transformer

Sep 2024Models: HTR-VT

TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models

Sep 2021Models: TrOCR

Recurrence-free unconstrained handwritten text recognition using gated fully convolutional network

Dec 2020Models: GFCN

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

Jun 2020Models: OrigamiNet-12, OrigamiNet-18, OrigamiNet-24

Other Optical Character Recognition Datasets

scut-ctw1500 mpqa cnn-/-daily-mail icdar2013 SROIE KITAB-Bench ThaiOCRBench PolEval 2021 OCR IMPACT-PSNC wikibio