Sample RankingRecallPrecisionHmean
A Simple Method for Key Information Extraction as Character-wise Classification with LSTM100.00%100.00%100.00%
Enetity detection100.00%100.00%100.00%
Bi-directional LSTM-CNNs-CRF (version2)100.00%100.00%100.00%
CLOVA OCR100.00%100.00%100.00%
H&H Lab100.00%100.00%100.00%
HeReceipt-withoutRM100.00%100.00%100.00%
GraphLayout100.00%100.00%100.00%
Ping An Property & Casualty Insurance Company100.00%100.00%100.00%
PATECH_CHENGDU_OCR100.00%100.00%100.00%
CITlab Argus Information Extraction (positional & line features, enhanced gt)100.00%100.00%100.00%
PATECH_CHENGDU_OCR_V2100.00%100.00%100.00%
PATech_AICenter100.00%100.00%100.00%
ASTRI-CCT-MSA100.00%100.00%100.00%
Character & Word BiLSTM Encoder100.00%100.00%100.00%
PICK-PAPCIC & XZMU100.00%100.00%100.00%
LayoutLM100.00%100.00%100.00%
LayoutLM + Excluding OCR Mismatch100.00%100.00%100.00%
BERT-MRC100.00%100.00%100.00%
HIK_OCR_Exclude_ocr_mismatch100.00%100.00%100.00%
SROIE LSTM - Axel Alejandro Ramos GarcĂ­a100.00%100.00%100.00%
End-to-end learning with PGN100.00%100.00%100.00%
Taikang Insurance Group Research Institute100.00%100.00%100.00%
Tencent Youtu100.00%100.00%100.00%
LayoutLM 2.0 (single model)100.00%100.00%100.00%
IE method100.00%100.00%100.00%
Applica.ai Lambert 2.0 + Excluding OCR Mismatch100.00%100.00%100.00%
Applica.ai Lambert 2.0 + Excluding OCR Errors + Fixing total entity100.00%100.00%100.00%
lstm standard method trained 100 epochs constant learning rate100.00%100.00%100.00%
Applica.ai TILT + Excluding OCR Errors + Fixing total entity100.00%100.00%100.00%
IE100.00%100.00%100.00%
Multimodal Transformer for Information Extraction100.00%100.00%100.00%
GEM AI -OCR TEAM100.00%100.00%100.00%
N100.00%100.00%100.00%
Linklogis_BeeAI100.00%100.00%100.00%
test excluding ocr error100.00%100.00%100.00%
layoutlm V2100.00%100.00%100.00%
Character-Aware CNN + Highway + BiLSTM 1.0100.00%100.00%100.00%
StrucTexT100.00%100.00%100.00%
1129-clm100.00%100.00%100.00%
1208-cblm100.00%100.00%100.00%
1208-cblm100.00%100.00%100.00%
LAKE100.00%100.00%100.00%
PIC-proj100.00%100.00%100.00%
CHL100.00%100.00%100.00%
GraphDoc100.00%100.00%100.00%
Character-Aware CNN + Highway + BiLSTM 2.0100.00%100.00%100.00%
GraphRevisedIE100.00%100.00%100.00%
loixc100.00%100.00%100.00%
plat-loixc-1g100.00%100.00%100.00%
Applica.ai roberta-base-2D75.00%100.00%85.71%
GIE75.00%100.00%85.71%
RoBERTa-base75.00%100.00%85.71%
RoBERTa-base finetuned on business documents75.00%100.00%85.71%
BERT with Multi-task Confidence Prediction75.00%75.00%75.00%
NER with spaCy model75.00%75.00%75.00%
With receipt framing75.00%75.00%75.00%
Location-aware BERT model for Text Information Extraction75.00%75.00%75.00%
BOE_IOT_AIBD75.00%75.00%75.00%
test75.00%75.00%75.00%
Cool Method Remix75.00%75.00%75.00%
great75.00%75.00%75.00%
Graph Neural Net with Bert Embeddings75.00%75.00%75.00%
Custom Named Entity Recognition75.00%75.00%75.00%
coldog75.00%75.00%75.00%
lstm deep75.00%75.00%75.00%
Qubitrics75.00%75.00%75.00%
Character Level BiLSTM75.00%75.00%75.00%
IE + GCN+ OCR75.00%75.00%75.00%
test275.00%75.00%75.00%
LSTM without OCRerror and RM75.00%75.00%75.00%
0 shots75.00%75.00%75.00%
plat-loixc75.00%75.00%75.00%
Key Information Extraction from Scanned Receipts50.00%50.00%50.00%
IFLYTEK-textNLP_v250.00%50.00%50.00%
SituTech_OCR50.00%50.00%50.00%
Token level multi modal bilstm50.00%50.00%50.00%
Token level bert embed + bilstm50.00%50.00%50.00%
2shot_tem0250.00%50.00%50.00%
5shout50.00%50.00%50.00%
layoutLMtest25.00%25.00%25.00%
LayoutLMv225.00%25.00%25.00%
Pytesseract + Character Level LSTM + Regex for Dates25.00%25.00%25.00%
1shot_position0.00%0.00%0.00%