cornelius785
join:2006-10-26
Worcester, MA
| reply to Kearnstd
Re: False Positives abound
i'd like to see their false positive rate and what happens when there is a false positive. i'd say it would be almost impossible to ensure there are no false positives. comparing binaries would be nice, but storage requirements would be insane (size and speed). binary comparing wouldn't catch images that have been slightly tweaked (slightly different resolution, watermark, a random line or 2, border, scaling, etc.) so a duplicate image processing technique would need to be used to find similar images. i sure would hope that before legal action is taken, a the result is checked to see if it is a false positive. |
