- Claim they are "beyond comparisons"…

- Refuse to let you perform your own comparative tests, with your cards and latents…

- Claim they don’t have a benchmark database to provide you with (5,000 as a minimum to 15,000 people to get proper statistics…). A 100-500 cards database is meaningless and can hide defaults which will only show up years after you bought a system.

- Companies that try to get you off the KEY subject: Where will my real HIT be on the list of candidates when my database grows. Companies that try to focus the conversation on speed rather accuracy are usually not very tempted by side by side tests in a reasonable size benchmark database… As far as we’re concerned speed is not the issue, all serious companies can add more processors (networked with your original System) each time you request a speed improvement. But does your real HIT stay high in the list of candidates as you grow the database, or will it fall down so far that you will spend hours on each latent review or just never find the real hit in the candidate list…

- Claim they have so many installed systems that you should not worry…Better performance does not always go with well established systems… Luckily most police Dept’s like to see and test what they are getting and are open to innovative ideas, so why not include some real comparisons in the process, and allow companies to show their new and innovative capabilities. We have experienced this before, in 1991 SPEX, was not known at all in the Forensics Light Source business, but is now THE worldwide leader, which sold large amounts of units to FBI (71 MCS-400’s, 10 CS-16’s), National Japanese Police (48 CS-16’s) and more recently French Gendarmerie (23 units) after these agencies compared carefully with the "established companies" through a lengthy cautious process…

- Fully automated palm entry: You can create a large database very quickly.

- Very powerful and accurate extraction for latents: Save a lot of time for Latent entry.

- Multiple parameters innovative technique using true Skeleton information for the searches.

- Fully Scalable concept: You can add more Matchers (CPU processors) to increase speed, Single PC, medium or large Servers, PrintQuest is fully expandable.

- No custom hardware: We use off the shelf Workstations (WS420) and Servers (6400) (all DELL brand for now, as they rate as the number one for reliability, while offering all components for the small PC Systems to the Large Servers, matchers for integration in racks, Raid-5 large storage systems…), so you are NOT locked with Custom costly hardware and unaffordable upgrades. (Models and brand may change without notice)

DID YOU KNOW THIS: PrintQuest is the ONLY low cost AFIS/APIS System with FBI certified "Scanner/Scanning software" tested for 500 dpi and 1000 dpi requirements.

DID YOU KNOW THIS ?

PRINTQUEST IS THE ONLY LOW COST AFIS SYSTEM WITH
FBI CERTIFIED “SCANNER/ SCANNING SOFTWARE” FOR TEN-PRINT
ENTRY (500 dpi) AND LATENT ENTRY (1000 dpi)

- Provide a PC already loaded with 5,000 to 15,000 cards (Verify it…), meaning you get 50,000 to 150,000 fingerprints and 10,000 to 30,000 palmprints to search YOUR cards/latents pairs through. Cards must be all different, not duplicated, for proper statistics: After a few searches, you will be able to verify it: If cards have been duplicated, they will show as consecutive cards in the candidate lists.

- Provide a ˝ day training/demo and then let one of your operators process the samples by himself. So you get a good idea of :

a/ How friendly the system is.

b/ How much the system does for you automatically, and how much remains to be done manually…Time counts…even if accuracy should be your first focus when you get your own small size Afis. Not all systems include powerful tools to avoid lengthy checking procedures

c/ How accurate systems are: Record the number of hits in top position in the candidate lists for each latent, the number of hits in position 2-5, 6-10 or worse. The best System is the one that provides the most hits in the positions closest to the top of the list, and with the least number of complete missed hits. The more latents you process, the more representative your test is.

Ideally the various vendors’ systems should be compared with the same minutiae extractions on the latents (but also track how much time is needed to reach this "perfect" minutiae extraction on the latents…

You should pull from your files:

We recommend that the latent fingerprints and palmprints range from 7-8 matching points only to 12-16 points, why make it easy and enter latents with 20 minutiae ?

If you have time, don’t limit your tests to "Latent to Ten-print" searches. Some brands do not feature the same accuracy for the "Ten-print to Latent" and "Latent to Latent" searches…

Test it all if possible. AFIS is not just a large investment that justifies getting the best, it MUST perform extremely well or criminals will remain at large.

Some of our customers have borrowed our evaluation systems for 3 weeks, after getting a 2-day training. We provided them with a 10,000 cards benchmark preloaded database (so-called background database), and they added their 500 cards, and 50 latents. We won’t mention (yet!) which famous brand they currently have, but their results with our system seemed to be the best they ever saw. When budgets come through for these systems, we will give updates.

For large systems decisions, we recommend that you enter a minimum of 250 of your cards and 50-100 latents. If you agree to release some results to other police agencies, and perform the tests in a reasonable amount of time, we will be happy to provide you with such evaluation system.

Some other Afis/Apis systems will show the real hits as candidate number 10 or 20 or…50. The lower the real hit in the candidate list, the MORE work you will have and less accuracy in the system you will get: A high accuracy system should show most latent to card hits (80-99%) in the position number one in the candidate list for database samples of 1,000 to 10,000, so you can be confident your people won’t spend their time reviewing long lists of candidates for EACH single latent you process and you can be sure you don’t miss matches (if they end up in position 25 or 50 or 100 in the candidate lists… ) Worse, some AFIS Systems will just completely miss on low quality latents (low number of points, poor images…).

Again, check the palmprint capabilities of all these companies that claim to have palm capability. Palm algorithms are very complex and we are confident the PrintQuest Solution will do the best job for you ! Performance with Fingerprints is NOT representative of palmprint performance. Check this hit:

Can other companies companies do this?