Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Identix (IDNX) -- Ignore unavailable to you. Want to Upgrade?

To: jean who wrote (12289)	1/27/1999 4:20:00 PM
From: David	Read Replies (1) \| Respond to of 26039

Interesting colloquy from a newsgroup (did you know IDT got its algorithm from Russia?): Subject: Att: Eugene Liunov From: "Sylvain Lemieux" <sylvain.l@sympatico.ca> Date: 1999/01/08 Message-ID: <w%nl2.630$uO3.4200@news20.bellglobal.com> Newsgroups: fr.sci.biometrie [More Headers] -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Thanks for the answer. You did not talk about Indenticator (www.identicator.com). I have bought their SDKs (about 20k $US). The speed is good (about 200,000 fingerprints/sec on a P400), but I haven't tested it yet on a bank bigger than 10,000 fingerprints. Their DFR-100 (512x512) is accurate, but need a frame grabber card. The new DFR-200 is frame grabberless but is only (256x256), so when I used the new 200 on fingerprint scanned for 10prints cards or from fingerprints captured with the old 100, the result are not very good. The SDK is suited for biometrics applications like access control, encryption and other application but I don't think that it is enough robust for AFIS system. I actually have a customer in South America who want to by our Mug Shot software (with facial geometry module for line-up), but he also want a full AFIS system. I don't know is the Identicator SDK is enough. Do you know them? When I meet with them in US, they claim that their algorithms as been develop in Russia, 10 years ago ??? I already know Printrak (since they have sold their system to Canada, Quebec province, and Montreal city police corp), but they are very expensive. Morpho as a good system (and a rolled-fingerprint reader), but this reader is far too expensive for my customer. Thanks, Sylvain __________________________________ Subject: Att: Sylvain From: "Eugene Lipunov" <lipunov@cityline.ru> Date: 1999/01/08 Message-ID: <915810017.297498@mao.cityline.ru> Newsgroups: fr.sci.biometrie [More Headers] Sylvain, Well, that's my failt. I know some about Identicator, but not so much. I have this DFR-200 OEM version, but I've never seen this 20k $ Identicator's SDK. My idea is to use this fingerprint scanner and it's device driver only, then provide the system with the own recognition algorithms. They have LS SDK (live scan SDK) with no recognition software that is much cheaper. And It's nice device for access control, etc., COMPAQ use this scanner to make fingerprint scanners under their trade mark and sell it for 100$ including NT Logon (GINA) software. It's hard to believe that Identicator's algorithms was developed in Russia 10 years ago. Maybe in Russia, but then not 10 years ago. Six years ago there was no AFISs in Russia at all. Ministry of Interior that time forced a competition between AFIS systems and they choose four AFISs that was recommended for ID devisions. Two systems was from Tcheliabinsk (city near Novosibirsk in Siberia), they developed "Papilon" and "SondaFres" AFISs. One was developed in some R&D center of Ministry of Interior - AFIS "Poisk". And one was "DactoPro" AFIS we developed. As I remember Identocator claims that they occupy 500 bytes per fingerprint, so you will feel comfortable even with million tenprint cards data base on one PC. They use minutae type algorithms, so I may assume that is fast enough, but require a good quality prints. If your customer is in US, it may require from you to correspond to some FBI standards, as I remember it is 500 dpi image resolution, inch by inch size for tenprint card images, 256 greyscale levels and WSQ image compression algorithm implementation, compatibility with other AFISs at least on image and demographic data exchange level. I see you have no answer about fingerprint recognition quality from Identicator itself. I do not know that either, but you have a hard way to check it. NIST provide developers with sample fingerprint data bases exactly for that purposes. You may try to get that via Internet or to buy a CDs, and that is raw format images. I know who may provide this dirty job, there are some projects in US for human identification on ATMs (that means "put your credit card into garbage") and this company have some testing methodics and data to compare fingerprint recognition algorithms quality for that projects, to provide false reject rate, false accept rate, etc. algorithm estimations for various types of data. They are on www.bessystems.com and they do some recognition sfotware too. They are in OldBridge, NJ, US by the way. Gene ________________ A few comments on this exchange: I don't believe that the IDT/Compaq product uses 500 bytes memory, but more like 100. I think "SDK" stands for "Software Development Kit." The Morpho product may be IDX or may be Morpho itself. "Gene" appears to be Russian, and Sylvain appears to be from Quebec province in Canada. If we do a little checking, we can probably figure out what French Canadian company has a "mug shot" product.