Ticket #36 (closed defect: fixed)
Example of image spam that gocr only finds gibberish in, even with pnm processing
| Reported by: | adam@… | Owned by: | decoder |
|---|---|---|---|
| Priority: | minor | Milestone: | |
| Component: | Image Analysis | Version: | |
| Keywords: | example-spam | Cc: | adam@… |
Description
Sorry if this is the wrong way to bring this up, but I couldn't find anything in the FAQ.gz, or Wiki, or IRC channel (or the mailing list, since it's private) that suggests what to do.
Here's an example of an image spam that gocr always interprets as gibberish, even after converting to a 3-color pnm and running "gocr -l 180 -d 2". I'm hoping that by posting this here we can somehow improve the system to detect it well. After all, if thwarting FuzzyOcr? is as simple as sending spam that looks like this...
By the way, why is the mailing list private? Is it to keep spammers from reading it? If so, isn't that like a very thin veil of security-by-obscurity?

