On 30/08/13 10:10, j.witvl...@mindef.nl wrote: > Some time ago i´ve been experimenting with fingerprints, and the result was > not encouraging... > From security point of view no that many problems (besides all well known > general issue´s with fingerprints). > I mean no false positive´s, but the huge amount of false-negatives: nine > times out of ten, I did not recognize correctly.
I've been using fingerprint-scanning for a year now over successive releases of Ubuntu on these XPS m1530 models (SGS Thompson reader) with libfprint. I've found it reliable. I think there are two things that lead to low false-negatives: 1) a good initial scan of the finger(s) 2) consistent conditions for reading For example, a usable but short scan or slightly contrived flexing of the finger during the initial scan will cause consistency problems later when the finger becomes more familiar with the action and changes the way it passes over the reader. Also, scans originally in bright light will not work well if reading is done in low-light conditions such as overnight when a room may only be lit by the LCD screen. I also found that an original scan done when the finger is moist will cause problems with reading when the finger is dry, dusty or dirty. The best conditions for the initial scan I've found are: 1. Do the scan in medium to low-light conditions 2. No excessively bright or directional lighting (avoid strong sunshine from windows, or desk-lamps) 3. Wash the finger(s) in hot water with soap and dry them well, which gives clean well-raised profiles 4. Glide the finger above the sensor with some flexing at the leading/trailing edge of the scan so the pad of the finger stays in contact with the centre of the reader I've also found that, for me, the middle and ring fingers of my right hand are more reliable than index or little. _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org https://lists.gnu.org/mailman/listinfo/grub-devel