Optical Character Recognition2020en

reuters-21578

Dataset from Papers With Code

Metrics:accuracy, cer, wer, f1
Current State of the Art

ApproxRepSet

Unknown

97.17

accuracy

Top Models Performance Comparison

Top 3 models ranked by accuracy

accuracy1ApproxRepSet97.2100.0%2REL-RWMD k-NN95.698.4%3Orthogonalized Soft VSM92.795.3%0%25%50%75%100%% of best
Best Score
97.2
Top Model
ApproxRepSet
Models Compared
3
Score Range
4.5

accuracyPrimary

f1

#ModelScorePaper / CodeDate
1
MAGNET
89.9
MAGNET: Multi-Label Text Classification using Attention-based Graph Neural NetworkCode
Feb 2020
2
VLAWE
89.3Feb 2019
3
KD-LSTMreg
88.9Apr 2019
4
LSTM-reg (single model)
87
Rethinking Complex Neural Network Architectures for Document ClassificationCode
Jun 2019
5
SCDV-MS
82.71Nov 2019

Related Papers6

Other Optical Character Recognition Datasets