A CNN-based OCR pipeline decoded timestamp characters on 7,760,911 scanned solar film images, reaching 97.9% image-level accuracy on tested images.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
astro-ph.SR 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Intelligent Recognition of Time Stamp Characters in Solar Scanned Images from Film
A CNN-based OCR pipeline decoded timestamp characters on 7,760,911 scanned solar film images, reaching 97.9% image-level accuracy on tested images.