Description
300 labeled images collected from Twiiter Arabic hashtags and used for fine-tuning a text loclization (pixel_link) part of OCR system. The datasets were used in our published paper "Detecting Spam Images with Embedded Arabic Text in Twitter".
External deposit with Mendeley Data.
External deposit with Mendeley Data.
Date made available | 8 Feb 2021 |
---|---|
Publisher | Mendeley Data |