← Search

Data · dataset · 2021

Synthetic datasets of adversarial images

Listed in Teesside University Research Data Repository

Description

We build synthetic datasets of images with embedded adversarial text to improve the robustness of OCR-based spam detectors. The datasets were used in our project (github.com/niddal-imam/Post-OCR-Correction).

Links

Where it is published

Catalogue records · 1

Topics

Inferred from text
Image 65% · Text 75%
Provenance · 1 source records, 9 field assertions
SourceKeyLast seenRaw
Teesside University Research Data Repositoryoai:data.mendeley.com/2g3c836mh3.118 h agoJSON v1
FieldAssertionExtractorEvidence
access_levelsource · researchdata tees ac ukconnector:researchdata_tees_ac_uk@1.0.0
concepts[field].local:field:earth-environmentalmapping · researchdata tees ac ukconnector:researchdata_tees_ac_uk@1.0.0
concepts[method].local:method:ocrmapping · researchdata tees ac ukvocabulary-mapper@1.0.0keywords['Optical Character Recognition']
concepts[modality].local:modality:imageenrichment · researchdata tees ac ukkeyword-concept-rules@1.0.0title+description (65%)
concepts[modality].local:modality:textenrichment · researchdata tees ac ukkeyword-concept-rules@1.0.0title+description (75%)
descriptionsource · researchdata tees ac ukconnector:researchdata_tees_ac_uk@1.0.0/metadata/dc/description
licensesource · researchdata tees ac ukconnector:researchdata_tees_ac_uk@1.0.0/metadata/dc/rights
publication_datesource · researchdata tees ac ukconnector:researchdata_tees_ac_uk@1.0.0
titlesource · researchdata tees ac ukconnector:researchdata_tees_ac_uk@1.0.0/metadata/dc/title