Requirements

The just command line tool https://github.com/casey/just

Raw Data Used in Paper

Crawler data

Warning: This data was collected automatically from a top list. Across website screenshots there is a lot of porn.

Top 1K (1.3 GB)

https://drive.google.com/file/d/1tQV7UarMKjv7cF1aSW4HKSrhKIeFikZh/view?usp=sharing

Top 10K

Incoming...

Top list (221 MB)

https://drive.google.com/file/d/1wo8uXII3PQmXO13rMwunSCBQc3QcspFJ/view?usp=share_link

Ground Truth Labled Data (17 KB)

https://drive.google.com/file/d/1e89O5qxPy8931kNITVMncUss350BuTzq/view?usp=sharing

Code and Tools

Crawler Data Collection

1. See the Quick Start in crawler/README.md.

Ground Truth Labeling

1. Navigate to the analysis/ground-truth-labeling directory.
2. Follow the instructions in the README.

Logo Detection

1. Navigate to the logo-detection directory.
2. Edit the justfile datadir variable to point to the crawler data folder.
3. Run just install-dependencies
4. Execute logo detection
   • just match-one: displays results for 1 image.
   • just match-all: Runs on a directory of login screenshots.
   • just match-parallel: parallelizes processing of match-all.