scooby / README.md
tsrdjan's picture
Update README.md
0974567
---
license: gpl-3.0
language:
- sr
- en
pipeline_tag: image-classification
tags:
- resume
- cv
- profile
- profile-page
- osint
- research
- crawling
---
# Scooby
Scooby is the first model created for the purpose of detecting profile pages while crawling.
It is trained mainly on scraped data from the sites of Serbian universities, but around 20%
of the data is scraped from websites of some organizations or companies.
## Preprocessing
For preprocessing, 2880x1620 resolution images were rescaled down to 360x480 (by mistake).
Number of channels is one, grayscale.