Milestone-Proposal talk:AFIS NEC
Advocates and reviewers will post their comments below. In addition, any IEEE member can sign in with their ETHW login (different from IEEE Single Sign On) and comment on the milestone proposal's accuracy or completeness as a form of public review.
Advocate's comments (posted on advocate's behalf by -- Administrator4 (talk) 14:09, 15 February 2019 (UTC)[edit | reply | new]
I have reviewed this proposal and have also done some in depth consultation discussions with image processing experts in the industry and academia that I had the opportunity to work with. And we agree that this effort deserves the recognition.
A few notes - this accomplishment is about building a high accuracy and a high utility system rather than a breakthrough invention of some fundamental technique or concept. So it is s little different from other milestones that typically recognize seminal efforts that later manifest into major fields of technology. The social impact of this milestone is huge - it enabled better crime solving and assisted in solving a few marqueee crime cases in the history such as LA’s Night Stalker.
Re: Advocate's comments (posted on advocate's behalf by -- Jason.k.hui (talk) 17:17, 23 February 2019 (UTC)[edit | reply | new]
As part of the Milestone process, please have two experts provide their reviews and approval of the Milestone in this forum. It is important that we have this documented.
Milestone Subcommittee Chair
I would suggest changing the word “criminals” to “suspects” in the last sentence and rephrasing to avoid vague passive voice: It enabled the world's police agencies to expedite searches for suspects, an efficiency that many public-safety experts valued.
Accuracy of fingerprint identification -- CSchlombs (talk) 03:09, 10 March 2019 (UTC)[edit | reply | new]
The accuracy of fingerprint identification has been called into question in recent years by studies of the National Academy of Sciences, the President's Councils of Advisors on Science and Technology, and others, for example:
Should this debate be taken into consideration in the formulation of the plaque citation, or at least be mentioned in the ETHW information?
Yes, I had the same concerns as Corinna, that there have been serious questions about the errors/careless use of fingerprints as evidence. Presumably, however, the plaque cannot get away with questioning this too much, both for word-length limits and what’s seen as acceptable. Moreover, one can say that the errors and abuses are in the human/social/institutional misinterpretation of the data (though then one can argue whether the technology itself promotes or facilitates such misinterpretation by giving an impression of perfect accuracy and making it so easy to search for “matches.” Anyway, it’s because I had the same concerns as Corinna that I suggested changing “criminals” to “suspects” and saying “many” public safety experts like this technology (leaving open the possibility that others may have reservations and that civil-liberties and privacy commenters may see things differently.
New Advocate for this Milestone Proposal -- John Vardalas (talk) 20:20, 14 March 2019 (UTC)[edit | reply | new]
I've been asked to be the new advocate for this proposal. I'm currently developing a list of external expert reviewers whom I will contact. Once these reviews are posted we can proceed with the evaluation of this proposal. In the meantime I am bringing myself up to speed on this proposal.
First external review of proposal -- John Vardalas (talk) 23:48, 22 April 2019 (UTC)[edit | reply | new]
Prof. Anil Jain, University Distinguished Professor, in the Dept. of Computer Science and Engineering, Michigan State University, has provided the first external review. He has had a distinguished career in Forensic Science. He was member of the United States Defense Science Board and Forensics Science Standards Board. He was Editor-in-Chief of the IEEE Transactions on Pattern Analysis and Machine Intelligence and is a Fellow of ACM, IEEE, AAAS, and SPIE. Jain is a Member of the U.S. National Academy of Engineering. Below is his evaluation quoted without any changes.
"I have read the Milestone proposal that you sent me. Below are my comments.
1. I know that NEC is a leader in AFIS technology and I have no doubt that they indeed built and installed the first latent AFIS in the world.
2. However, it would be nice to see explicit documentation of the claim that NEC was the first company. Some of the links that have been provided in the proposal are either broken or do not have any dates.
For example, while I can read the NEC brochure
A3: NEC AFIS Brochure, NEC 1986, I could not locate any date printed on the brochure itself.
Similarly, I could not open the link
N3: TIME No. 41 October 14, 1985 (p53) (TIME1985Oct P53.pdf)
I would suggest that they copy the socue and create a separate file and provide the link.
I a few instances, the proposal provides a link to an entire book, rather than excerpting a few sentences from the book to make the point.
3. In the early 1980s, and even now, there were less than 5 companies selling AFIS. It would be useful if the proposal provides when these other companies built competing systems. That will further bolster NEC's claim.
4. If this supporting information can be documented, I am happy to approve the proposal. NEC deserved this recognition.
5. Wording of the Plaque Citation is accurate.
6. The proposed milestone represent a significant technical achievement."
I appreciate Prof. Jain's review and his valuable suggestions.
As to the printed date of NEC AFIS Brochure is shown in the codes on the last page, bottom right. "8607300001" is supposed to mean that this Brochure is the 1st printing in July 30, 1986.
As to the absence of the supporting document, the copy of those document are now linked from the web page. Thank to Dr. Vardalas and the ETHW administrator for providing the link.
As to the date when the other companies built the competing systems, relevant information are provided in page 6-3 through page 6-9 of the Fingerprint Sourcebook (N2) and in page 5 of CAL-DOJ Status Report (N4). According to these document, Printrak's Latent AFIS systems installed in mid-1970's were not able to meet the large-scale operational requirements.
Thank all again.
Re: Re: First external review of proposal -- John Vardalas (talk) 21:50, 27 April 2019 (UTC)[edit | reply | new]
I've notified Prof. Jain that you have responded. I'm waiting for his reply. We are also waiting another external review to come in.
In the late 1970’s as an Inspector with the San Francisco Police Department, I was tasked with research into the new Rockwell Printrak Automated Fingerprint Identification system and the FBI Finder system upon which it was based. I travelled to all installed sites to examine their operations. In 1982, San Francisco went out for a competitive bid for an A.F.I.S even though at the time only the Printrak system was being actively marketed. Surprisingly, we received bids from an English firm, Logica, and from N.E.C. Since our primary focus was on the crime solving function, we devised a comprehensive latent print benchmark test which we administered to each of the three systems. Logica and NEC passed the test, but because only NEC had a current large system in operation, NEC AFIS was selected.
I have read the NEC Milestone proposal and find that the claims made therein are accurate and without exaggeration.
NEC built the first large scale AFIS capable of identifying prints from crime scenes with a high degree of accuracy and success.
The F.B.I.’s Finder System was larger but was built primarily for use in technical searches of newly arrested persons against its existing data base. It was not effective in searching latent prints, a point that the Bureau readily admitted. The fact that early Printrak AFIS models used the same algorithm explains why those other systems in use at the time also suffered from poor latent print accuracy. (Consumers Guide to Fingerprint Systems, Ident. News, 1986)
Tracking AFIS identifications through the criminal justice system in San Francisco proved that the system had a significant impact on lowering the burglary rate during the first five years of operation while during same period burglaries increased in other cities with similar demographics. (In the U.S., approximately 70% of latent fingerprints are taken from burglary scenes.) (Fingerprint Sourcebook (N2 6.6-9)
There is no question that NEC set the bar at a high level for the competition, and companies like Rockwell, Morpho, Cogent, and Lockheed Martin later rose to the challenge by substantially improving their own fingerprint systems thereby making the communities served by their customers safer.
For their historic and pioneering efforts, I endorse NEC’s Milestone application.
Re: Review by Kenneth Moses -- John Vardalas (talk) 17:22, 30 April 2019 (UTC)[edit | reply | new]
Kenneth Moses has over forty years of experience in forensic evidence. He established the Crime Scene Investigations Unit of the San Francisco Police Crime Laboratory in 1983 and was instrumental in the installation of automated fingerprint systems in San Francisco and at the California Department of Justice. Over the years, he has worked with attorneys, with local, state, and federal agencies, and with private companies in a number of complex investigations. Mr. Moses holds an interim top-secret security clearance for work with the Office of Military Commissions in crimes involving members of the armed forces. He has been active in national efforts to establish professional standards in forensic science. Mr. Moses currently serves in as Laboratory Director of Forensic Identification Services in San Francisco.