Die Biometrie toter Augen

Gepostet vor 9 Monaten, 22 Tagen in #Science #Tech #Biometrics #Death #Eyes #Security

Share: Twitter Facebook Mail

Mjc5MDM4Ng

Samsung hat gestern sein neues Galaxy Note 7 mit Iris Scanner vorgestellt. Sehr viel interessanter als neue Telefone aber ist die Arbeit von Adam Czajka und Piotr Maciejewicz, die die Biometrie toter Augen erforscht haben. Die haben nun rausgefunden, dass Augen noch 5 Stunden nach dem Tod erkannt wurden, in Einzelfällen bis zu 27 Stunden.

eye2

They used both infrared and visible light cameras to collect images, getting the first set at 5 hours after death, and doing two other sessions at about 16 and 27 hours after death. Czajka and PhD student Mateusz Trokielewicz then tested whether several commercial products and one open-source iris matching system worked on the dead eyes.

He found that the dead eyes were easily recognizable in the first session, and some could still be used for identification as late as 27 hours after the person’s demise. While the cornea gradually became cloudy over time as the corneal cells broke down, that opaqueness didn’t interfere with the infrared imaging. Czajka’s now conducting research using eyes that are several days or even weeks past their expiration date.

Paper: Post-mortem Human Iris Recognition, aus den Conclusions:

This paper presents the only study that we are aware of regarding post-mortem use of human iris as a biometric identifier. Contrary to claims common in the biometric community, our results show that human iris can be successfully employed for biometric authentication for a number of hours after death. Empirical study incorporating four different iris recognition method has shown that a significant portion of irises can be successfully recognized 5-7 hours after a person’s demise (with FNMRs equaling 0% to 8.33% for the best and the worst performing method, respectively).

However, this percentage is expected to decrease significantly as time period since death progresses, reaching FNMRs of 26.67% to as much as 86.67%. Comprehensive medical commentary is also presented to explain the underlying causes of such behavior, with processes assorted with corneal opacification, drying and loss of intraocular pressure recognized as the most probable sources of recognition errors.

Second Life is full of dead Bunnies

Ihr erinnert Euch an Second Life? Eine Mischung aus Game, virtueller Community und rudimentären Social Dings? Dort konnte man anscheinend…

A History of Human Skull Drinking Cups

Die beste Youtuberin des Todes über Trinkbehältnisse aus fertigverwesten Köpfen: Here's what my other Fav of Death (Lindsay Fitzharris) had…

How To be sad

„There is such pressure on us to be cheerful. And yet in order properly to be content, we need to…

Bambi eats human Bones

Forensiker haben zum ersten mal einen Hirsch (oder zwei, sie wissen nicht, ob es derselbe ist) beim Kauen auf Menschenknochen…

Fuck Zombie-Ants. Here's eyeball-parasite-controlled Zombie-Fish!

Scheiß auf pilzgesteuerte Zombie-Ameisen, es gibt augenparasitgesteuerte Zombie-Fische! Der Parasit in diesem Fischaugenfall ist nochmal ein wenig abgefahrener, als der…

Creatives see different

Wissenschaftler haben zum erstenmal gemessen, dass kreative Menschen (die vor allem das Persönlichkeitsmerkmal Openness aufweisen) „physisch anders“ sehen, als der…

Nathan Reidts Vieraugen

Tolle Vieraugen von Herrn Nathan Reidt:

Tetrachromatische Super-Vision per Farbfilter

Vor ein paar Jahren hatte man in England die erste Tetrachromatin der Welt gefunden, eine Frau mit vier unterschiedlichen Farb-Rezeptoren,…

Photos from the Body Farm

Großartige, teilweise sehr verstörende Bilder von Robert Shults von einer Body Farm in Texas, dem größten Outdoor-Verwesungs-Labor der Welt. Die…

Excellent Eye-Eater

I'd vote for her. Eye-Eater Jenny for Prez 2048! (Hab ich früher auch manchmal gemacht. Eat more eyeballs!)

Dude crushed by 6-Ton-Porn-Stash (UPDATE: Fake.)

[update] Because ofcourse: That Viral Story About a Japanese Man Crushed to Death by His Porn Collection is Totally Fake:…