Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Download
Standard view
Full view
of .
Look up keyword
Like this
2Activity
0 of .
Results for:
No results containing your search query
P. 1
Morph of Ace Investigate PCSD CaseStudy.snglpgs[1]

Morph of Ace Investigate PCSD CaseStudy.snglpgs[1]

Ratings: (0)|Views: 93|Likes:
Published by Brackett427
This is a case study by MorphoTrak, Inc. of the use of its MFI (MorphoFace Investigate) technology by the Pierce County (WA) Sheriff's Department, c.2009.
This is a case study by MorphoTrak, Inc. of the use of its MFI (MorphoFace Investigate) technology by the Pierce County (WA) Sheriff's Department, c.2009.

More info:

Categories:Types, Brochures
Published by: Brackett427 on Jul 29, 2009
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

04/07/2015

pdf

text

original

 
Case Study 
Pierce County Sheriff Department
Pierce County Sheri Speeds Bookings and SolvesCrime with Automated Face Recognition System
Faced with sta reductions and budget cuts, the Pierce County Sheri’sDepartment (PCSD) in Washington State initiated a pilot in mid-2008 toexamine the use o automated ace recognition to enhance the eciency o its orensics operations. PCSD rst applied the technology to minimize themanual work involved in its booking process and ree up time or orensic technicians to support criminal investigations. Based on the success o the trial, PCSD determined that by adding ace recognition to ngerprint identication in a ully automated ‘lights-out’ workow, it may ultimately save80 percent o the staf time required to process and validate the identities o repeat ofenders. In September, PCSD expanded use o the technology intocriminal investigative work on an identity thet case. In its rst application,automated ace recognition matched a grainy ATM photo o a suspect withan archived mug shot, allowing detectives to make an arrest in the case.
 
matching is considered the initial phase o the identication process; while the mug shot is usedas the secondary identier enabling orensic technicians to veriy ngerprint results.Prior to the pilot, orensic technicians manually retrieved records or repeat oenders romthe PCSD criminal record database (CHRI) as identied through ngerprinting. The technicianthen visually compared the new mug shot with the old one to positively conrm the suspect’sidentity. I both ngerprints and mug shots point to the same identity, it is considered conclusive.The CHRI would then be updated with new prints, photo, and booking details, and ultimately beorwarded electronically to the state and the FBI or inclusion in their respective databases.
Speeding The Booking Process
Going into the 2008 pilot, PCSD’s Wilkins estimated that up to 85 percent o suspect bookingsin a given day involved repeat oenders, i.e., criminals who have been arrested and processedby the Sheri’s Department previously. As repeat oenders, the vast majority have both theirtenprints and mug shots in PCSD databases. The department began storing digital photographso suspects in 1992, amassing a database o more than 500,000 mug shots today.During the typical booking process, the suspect is ngerprinted and photographed. The tenprintsare input into the Sagem Morpho AFIS as part o the identication process to determine i theindividual is already in the database with a history o prior arrests and convictions. Fingerprint
 MorphoFace Investigate is a robust and scalable acial recognition application including case evidence management,biometric matching, and orensic evaluation tools or intelligence analysis and investigative crime solving tasks.Based upon Sagem Morpho’s award winning, highly-scalable MetaMorpho AFIS architecture, MorphoFace Investigate can accept remote submission o acial images rom mobile devices such as camera equipped cell-phones and digital cameras connected to squad car laptops,enabling law enorcement to quickly identiy the potential use o aliases and suspect identities while in the ield.
Automated ace recognition is a biometric technology that only recently reached the level o maturity where it can be deployed operationally in security and lawenorcement environments or the purposes o identiication and investigation. Like ingerprint identiication technology, acial recognition can recognize uniqueindividuals by using one image to search against a database o enrolled images. But the similarity between ace and ingerprint recognition ends there. Whileingerprint algorithms utilize eatures which remain stable over time, such as ridge endings, biurcations and skin pores, ace recognition algorithms must be armore sophisticated because the appearance o aces can change drastically with age, emotion, presence o acial hair and even the level o sobriety and health.
PCSD - An Early Adopter o Technologies
Pierce County Sheri’s Department has a tradition o being an early adopter o automated tools andtechnologies – including biometric identiication – in the ight against crime. In July 1986, the department became theirst law enorcement agency in the country to purchase an Automated Fingerprint Identiication System (AFIS) rom Sagem Morpho,headquartered in Tacoma, Washington. This technology later ormed the basis o the FBI’s current IAFIS system, which today houses more than70 million ingerprint records, and in early 2009 processed a record 163,000 ingerprint searches in a single day.PCSD serves a population o more than 800,000 in an area encompassing the southern portion o the Seattle – Tacoma Metroplex. Like commercial businesses and otherlaw enorcement agencies around the nation, PCSD is continually expected to nd more cost-eective ways to deliver services during a time o reduced budgets and increasedoperational demands.Faced with these challenges, Wilkins looked or ways his department could utilize emerging technologies to reduce the timerequired to veriy the identities o suspects during the booking process. He reasoned that by eliminating some o the time-consuming manual eorts involved in this process, his group o orensic technicians could better support their ellow oicers,as well as the public at large, by applying more time to criminal investigations.One such promising technology was MorphoFace Investigate (MFI), an automated ace recognition biometric system beingintroduced by Sagem Morpho.
 
“The budget shrinks every year,making it dicult to replace people or enlarge the sta when necessary,” said SteveWilkins, PCSD Forensic ServicesManager.
Facial recognition solution or law enorcement agencies
Case Study 
Pierce County Sheriff Department
 
Finding Suspects Faster
Traditionally, the selection o candidate mugshot records depended on developingone or more suspects in a case. A laborintensive, and usually impractical,alternative was to select a larger groupo mug shots consisting o persons pos-sibly involved in the crime (given the type o oense, geographical location and other investigativeactors). In the extreme, an entire, set o records mightbe searched in cases o suiciently high priority. Whenrecords are selected in this ashion the comparisonprocess is reerred to as a ‘cold search’. Until the com-puterization o these records, such cold searches couldnot be routinely perormed.Face identication, or the irreutable determination o an identity, cannot be perormed solely by a computerusing technology available today – the subtleties o theprocess, especially when one considers the wide rangeo images available rom modern day crime scenes,make it dependent on expert opinion. Still, automatedace recognition has a dramatically eective auxiliaryrole o selectively retrieving candidate records rom largedatabases. This capability allows cold searches to beconducted routinely, and has proven to provide investigatorswith candidate pools o suciently high condence thatmany crimes are now being solved where continuedinvestigation was not previously possible.
Solving Crimes withFace Recognition
In September 2008, PCSDexpanded the application o MFI beyond booking suspectsto investigating crimes. In onesuch case, detectives were investigating a serieso identity thets in which the suspect was stealing ATM cardsand then using them to withdraw money rom the victims’ bank accounts. Unortunately, the only evidence available weregrainy photographs acquired rom ATM cameras where the perpetrator had made raudulent withdrawals o cash.A detective working the case thought he recognized the perpetrator but didn’t know the individual’s name. Fourphotographs rom our separate thets were sent to the orensics laboratory where Wilkins used the MFI system tocompare them against the mug shot database. Despite the bad camera angles and low image resolution, the systemreturned two mugshots belonging to one suspect, in a matter o seconds.The detectives reviewed the records o the suspect and learned thatshe had a past history o identity thet. Armed with this inormation,the investigators presented the evidence to a Superior Court Judge.Although the Judge had not seen acial recognition evidence in thepast, based on the strength o the algorithm evaluations, the strongsimilarity in appearance, and a previous record o identity thetrelated crimes, she issued an arrest and search warrant. In October 2008, detectives conronted the suspect and searched herresidence. A cache o additional evidence linking the suspect to numerous cases o identity thet was ound.Not only were ATM cards ound or the known cases o ATM thet, but detectives discovered additional materials, such ascredit cards, bank statements, bank account and routing numbers, pertaining to dozens o other individuals that couldeventually have been used to perpetrate additional crimes. The suspect plead guilty to 11 crimes in connection with theATM thets and was sentenced to 9 ½ years.
“From the time I received the email [with the attached  photos], the process took about 15 minutes,” said Wilkins.
This manual process takes hours o sta time each day, explains Wilkins. Not only could theorensic technicians’ time be better spent on more critical tasks, but the delay in updatingcriminal les slows investigations.“My three technicians work Monday through Friday rom 8 am to 5 pm,” said Wilkins, addingthat details o an arrest on Friday night won’t make it into the departmental, state and FBIdatabases until Monday at the earliest.To evaluate the accuracy o Sagem Morpho’s ace recognition algorithms, PCSD began using mugshots o known repeat oenders rom each day’s bookings as a source or searches against thedatabase o nearly 500,000 images. Without ail, the system routinely returned matches withinseconds. Ater more than 10,000 searches, the system correctly matched new mug shots withmug shots rom the individual’s previous arrests an impressive 94.4 percent o the time.What makes the 94.4 percent success rate impressive is the act that many o thematched mug shots were taken 10 to 15 years ago. Needless to say, acial eatureschange considerably over such a period. Wrinkles and changes in acial hair, weightgains or the dramatic weight loss a
 
ttributedto meth addiction, and widely varyingemotional states, can make current acialappearances dier greatly rom those o a decade ago. Despite these issues, the MFIacial recognition algorithms repeatedly ound correct matches.Wilkins estimates that by ullyautomating the booking process –or making it ‘lights-out’ – sta timewill be reduced by 80 percent onarrests involving repeat oenders. Identication is already a semi-automated process usingngerprint technology at PCSD, but with the integration o automated ace recognitioninto the workfow, a ully autonomous process is expected. Once ngerprints and mugshots are taken during the booking process, the identication o repeat oenders as well asthe updating and distribution o criminal les will occur automatically with no additionalmanual intervention – the task will be completed within minutes, rather than days.The secret behind MFI’s high accuracy lies in the intensive development and researchbehind its algorithms, and its specic ocus on workfows related to law enorcement,deense, and intelligence identication tasks. Sagem Morpho and its parent companySagem Sécurité are involved in biometric activities worldwide, providing unparalleledopportunities to work collaboratively with academia, law enorcement, deense andcommercial interests in the advancement o biometric technologies.
“In my opinion, [the system] has passed with fying colors,”  said Wilkins.

Activity (2)

You've already reviewed this. Edit your review.
1 thousand reads
1 hundred reads

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->