Talk:E-FIT

John Platten
I'm the John Platten that wrote the software so have enough materials accumulated along the way to un-stub this to a full article. The question is what do other people want to hear about? Leave me a message here and I'll see to it sooner or later. My fear so far has actually been to overdo my contribution and break Wikipedia rules in some way. But the article is classified as a stub now which I guess is an invitation to put a bit more... Opinions welcome on what that should be.

John Platten 23:44, 8 February 2007 (UTC)

I think the article could definitely use a bit more on the history/development of the software; perhaps some information on the scale to which it is used internationally, and different methods/techniques involved; some cases in which the software has been successful/unsuccessful; and, finally, the 'Research' section is a mess and doesn't make much sense. (Some pictures of the software in use would be great as well if anyone has access to them and permission to upload them.)

193.188.105.220 (talk) 21:35, 23 October 2008 (UTC)

Hi John, I think this is a great start and would welcome further information about E-FIT and the developments of the system. I'm aware that EFIT-V is the new standard and an explanation of that would work well here. The research that you identified were some of the best in this field but there are lots more which could help people who have an interest in this field. You have my vote. —Preceding unsigned comment added by 203.97.217.36 (talk) 11:28, 14 April 2009 (UTC)

1984 = Early use of "e-"? ...and other notes
Article says: "The term itself was proposed by Janina Kaminska at the UK Home Office in 1984. It falls on the leading edge of or predates the general use of "e" for electronic and retains the "E-" it started with to this date." -- I see that this is not cited, and may not be correct, as I see posts from 1981 in Google Groups using the term "email". -- 201.37.230.43 (talk) 22:22, 9 August 2009 (UTC)

Well John, I wrote the initial code and handed the project over to you! (Clive Higginson)on an apricot xen (precursor to the xen-i which was the IBM compatible version of the xen) system with an external I/O systems framestore (pluto?) if I remember correctly, earlier versions of the system (no code relation at all) were on an ICL mainframe using paintbox software/hardware. —Preceding unsigned comment added by 86.139.75.216 (talk) 10:41, 3 October 2010 (UTC)

As for the "E" prefix, we were (the SRDB of the Home Office where this project was based) already using email extensively within the department based on the Data General CEO office automation system on an MV6000 so "E" definitely predates efit even within the organisation, I even remember discussions about what to call the project and I am pretty sure Janina did indeed come up with the name but "e" was already in pretty general use at the time. —Preceding unsigned comment added by 86.139.75.216 (talk) 10:47, 3 October 2010 (UTC)

Also Tony Kitson was involved in the project long before myself or you John, so He does deserve some sort of mention and comment ;-) —Preceding unsigned comment added by 86.139.75.216 (talk) 11:01, 3 October 2010 (UTC)

Just to note that all Clive's comments are accurate and it's good to hear from him. The project was initiated by Tony Kitson and prototypes written by Clive and Mark Frampton. However... I did write the software for 16 years and took it well beyond any form of research project into a commercial project sold in multiple countries and that really was pretty hard work. I'm happy to call Tony and Clive "the fathers" of E-FIT in the same sense as that would be used of any discovery or invention. The short article also overlooks the work of John and Jean Shepherd at Aberdeen University in devising the descriptive index and Peter Bennett, former head of the Met's forensic art team who purchased the E-FIT software from the first commercial team and ran the project for ten (?) long years originally sustaining it through just picture database releases until I came back latterly to first perform a few maintenance releases and then the final rewrite in C# under contract. The software is now tended by Matthew Maylin of VisionMetric who is also the author of E-FITv, which is a ground up re-imagining and not on the same code base or with the same path from witness testimony to image. (To answer another poster - the code is private commercial and held by VisionMetric).

No mention of Peter would be quite complete without the role he played in emphasizing that end to end professionalism is the key to results. Elsewhere in the discussion page we read how a specialist team's results collapsed when the task passed to general scenes of crime officers. It has in fact been clear for a very long time that the interview and its integration with the composite, the path taken through construction and the quality of the witness are all vital.

Professional teams often refuse to even attempt a composite in situation such as a passerby having someone just brush past them. I mean - could you remember that? Where the system has proved reliable many times is in crimes of personal violence, especially rape, and in the combination of an E-FIT providing a name, often from another police officer who recognises a suspect and DNA evidence for the conviction.

With regard to lab tests of effectiveness. Really - unless you are actually going to rape the test subject I don't think it would achieve anywhere near the degree of imprinting that takes place. A suspect who achieves an accurate E-FIT of someone who haunts their dreams nightly is not so unexpected. All I can tell you is Stratclyde Police solved a murder in which they had no other immediate lead when they were just trialing the software, Dave Parker of Greater Manchester police relates one case in which the perpetrator's brother dragged him into the police station having seen his face on BBC Crimewatch and numerous other anecdotes from serving officers point to similar success - if the imprinting of the target face was heavy enough - and the operator uses other expertise alongside the system itself. John Platten. 84.92.230.173 (talk) 03:50, 29 December 2012 (UTC)

Question
So is this a variety of software (like "word processor"), or a specific software program (like "Word" or "WordPerfect")? If the latter, is E-FIT a proprietary &/or commercial program, or is it freeware or Open source? (Can anyone download it?) The article is unclear about these matters. -- llywrch (talk) 15:08, 28 April 2011 (UTC)
 * I had assumed that it was entirely owned by the Home Office and thus not available, in any form, in the public domain (which makes it an unusual topic for a Wikipedia article?) But like you, I await informed expert advice. Martinevans123 (talk) 22:45, 28 April 2011 (UTC)

ANSWER - The Home Office issued a sole license to Aspley Limited to develop the software beyond a research project and I left to work on it in the private sector at that point, so no, there is no public code. Tony Kitson (answers above) renaged on this and sent a letter of which I still have a copy, stating that the Home Office "Was now no longer prepared to offer a sole license" (note 'no longer') and the code was also given to a company called Noble Campion who wrote CD-FIT, later Pro-FIT. This so annoyed me at the time, along with comments made to a group of police officers by that company that I worked night and day on creating an E-FIT that would finish them in the market place, our Managing Director at that time being of the opinion that it was a better route than legal action (so clearly what I am saying has not been through court). This leaves two commercial UK based systems from the original project and no public license. Note that we are talking 1880's here and the code is pretty much a Xen and a Pluto as Clive Higginson says above - so a public version would not do anyone any good anyway. As Home Office research rather than a live concern its lost in the mists of time. Company policy when we had the E-FIT license as Aspley Limited was to sell copies to Police agencies and other professional parties such as University Psychology departments only to avoid contamination of the police use of the images in the public's mind. You would have to ask current owner VisionMetric directly about their policies. The software is unlikely to be cheap, as such, at any point as its a fairly vertical market. John Platten 84.92.230.173 (talk) 04:09, 29 December 2012 (UTC)

is this written in English???

ANSWER - It was written in English and then translated to the many languages in which it now runs. There were about 7 at the last count on my watch but VisionMetric would know better if this has been added to subsequently. John Platten 84.92.230.173 (talk) 04:09, 29 December 2012 (UTC)

"During the early 1990s Greater Manchester Police's dedicated Facial Identification Unit recored over 55% success rate with the E-FITs created by their specialist officers.

It was later disbanded and the duties were handed over to the forces Scenes of Crime Officers. !!!! Who were expected to compile E-FIT composites in between taking scrime scene photogrpahs.!!!!! The success rate fell dramatically." — Preceding unsigned comment added by 89.43.195.27 (talk) 23:44, 16 March 2012 (UTC)

ANSWER - That is a great comment that I have more fully answered above. The specialist user was Dave Parker and I personally would rank him well among the top ten operators. Few of his peers would disagree with that or with the assertion that the original E-FIT required a fair degree of skill to operate - primarily on the interview and process side - the software itself has never been that difficult: I once trained an actor for BBC's THe Bill crime show in just twenty minutes, to move some things around - but that does not mean it would have been a wise idea to let him loose on real crime victims. John Platten 84.92.230.173 (talk) 04:09, 29 December 2012 (UTC)