Is Forensic Speaker Recognition The Next "Fingerprint?"

Take a fingerprint... for that matter, go ahead and take a palm print. Now, take a voiceprint. In this day and age, forensic biometric analysis is extraordinarily complex. In a world where we analyze everything from irises to earlobes, what can science tell us about voice?

One increasingly popular form of analysis is forensic speaker recognition (aka voice biometrics or biometric acoustics). Forensic speaker recognition (FSR) has unequivocal potential as a supplementary analytic methodology, with applications in both the fields of law enforcement and counterterrorism (for details, see the last section of the 2012 book on FSR Applications to Law Enforcement and Counter-terrorism).

The utility of the FSR process is either one of identification (1:N or N:1) or verification (1:1).

• 1:N Identification -- Imagine you have a recording of a voice making threats over the phone. The speaker identification process allows you to query a database of acoustic recordings of known suspects for comparison against your target voice to identify more threats he/she might have made.
• N:1 Identification -- Imagine you have a bunch of voice recordings and you want to know in which of them, if any, a certain speaker participates. 
• 1:1 Verification -- Imagine you wish to grant someone access to a building or secure location by assessing whether or not they are who they say they are (this aspect of speaker recognition is less applicable to analysis and more applicable to security). 

That said, the CIA, the NSA and the Swiss IDIAP all turned to automatic speaker verification systems in 2003 to analyze the so-called Osama tapes (for details of the approach, see Graphing the Voice of Terror). This case provides an excellent opportunity to note the distinction between automatic speaker recognition performed by an algorithmic machine and aural speaker recognition performed by acoustic experts. 

The cornerstone methodology supporting forensic speaker recognition is voiceprint analysis,or spectrographic analysis, a process that visually displays the acoustic signal of a voice as a function of time (seconds or milliseconds) and frequency (hertz) such that all components are visible (formants, harmonics, fundamental frequency, etc.).

(Note:  For those who are more acoustically inclined and would enjoy a well-written read on all things acoustic from military strategy to frog communication, Seth Horowitz's new book The Universal Sense: How Hearing Shapes the Mind comes with my highest recommendation.)

Spectrographic analysis differs from human speaker recognition in that it provides a more quantifiable comparison between two speech signals. Under favorable conditions, both approaches yield favorable results: 85 percent identification accuracy (McGehee 1937), 96 percent accuracy (Epsy-Wilson 2006), 98 percent accuracy (Clifford 1980), 100 percent accuracy (Bricker and Pruzansky 1966). These approaches, however, do not come without caveats.

Forensic speaker recognition has many limitations and is currently inadmissible in federal court as expert testimony. Bonastre et al (2003) summarize these limitations quite well:  

"The term voiceprint gives the false impression that voice has characteristics that are as unique and reliable as fingerprints... this is absolutely not the case."

The thing about voices is that they are susceptible to a myriad of external factors such as psychological/emotional state, age, health, weather... the list goes on. From an application standpoint, the most prominent of these factors is intentional vocal disguise. There are a number of things people can intentionally do to their voices to drastically reduce the ability of machine or human expert to identify their voice correctly (you would be amazed at how difficult it is - nearly impossible - to identify a whispered voice). Under these conditions, identification accuracy falls to 40 - 52 percent (Thompson 1987), 36 percent (Andruski 2007), 26 percent (Clifford 1980). 

Top: Osama bin Laden's "dirty" 2003 telephonic spectrogram
Bottom: Osama bin Laden's "clean" spectrogram
Source: Owl Investigations

More problematic still is communication by telephone. Much of the input law enforcement and national security analysts have to work with comes from telephone wiretaps or calls made from jail cells. Telephones, cellphones in particular, create a filtering phenomenon of an acoustic signal, whereby all acoustic information under a certain frequency simply does not get transmitted (within this frequency range lie some of the key characteristics for voice identification). 

While the forensic speaker recognition capability has come a long way since 2003, the consensus among the analytic community remains that it is not a stand-alone methodology, rather a promising supplementary tool. Biometric analysis was also a topic brought to the Intelligence Technology panel of the 2013 Global Intelligence Forum conference this year. Of note was the expanding applicability and increasing capabilities of all biometric technologies. 

Thus far, the Spanish Guardia Civil is the only law enforcement agency worldwide to have a fully-operational acoustic biometric system (called SAIVOX, the Automatic System for the Identification of Voices). In the Spanish booking process, just like we take fingerprints, they take voice samples that they then contribute to a corpus of over 3,500 samples linked with well-known criminals and certain types of crime. 

In 2011, the FBI commissioned NIST to launch a program on "investigatory voice biometrics." The goal of the committee is to develop best practices and collection standards to launch an operational voice biometric system with robust enough corpora so as to serve as a useful tool in ongoing investigations, modeled off the Spanish system. (This is an ongoing project and you can read the full report here).

FSR is not a perfect methodology, but one that can add substantial value on a case-by-case basis. It is of high interest to the US national security and law enforcement analytic communities.

Additional reading:

Andruski, J., Brugnone, N., & Meyers, A. (2007). Identifying disguised voices through speakers' vocal pitches and formants. 153rd ASA meeting.

Bonastre, J. F., Bimbot, F., Boe, L. J., Campbell, J. P., Reynolds, D. A., & Magrin-Chagnolleau, I. (2003). Person authentication by voice: A need for caution. Eurospeech 2003.

Bricker, P.D., & Pruzansky, S. (1966). Effects of stimulus content and duration ontalk identification. The acoustical society of the Americas, 40, 1441-1449.

Clifford, B. R. (1980). Voice identification by human listeners: On earwitnessreliability. Law and human behavior, 4(4), 373-394.

Epsy-Wilson, C. Y., Manocha, S., & Vishnubhotla, S. (2006). A new set of features fortext-independent speaker identification.

McGehee, F. (1937). The reliability of the identification of the human voice. Journal of general psychology, 31, 53-65.

Parmar, P. (2012). Voice fingerprinting: Avery important tool against crime. J Indian academy forensic med.,34(1), 70-73. doi: 0971-0973

What Is Crowdfunding And Why Is It Important To Intelligence? (ENTINT)

http://widget.launchrock.com/

Six more days...

If all goes well, in six days, I intend to launch my first game, Widget, on Kickstarter  (if you are interested in the game itself, you can get more info by clicking on this link).  Once launched, I will have 30 days to get "funded" by various "backers".  If I fail to reach my pre-designated goal, I get nothing at all.

That's how crowdfunding works.

Well, at least, that is how Kickstarter works.  Kickstarter is the oldest and most popular crowdfunding platform currently available.  For those of you unfamiliar with these platforms, you probably should be.  This is not just for entrepreneurs or people interested in entrepreneurial intelligence, either.  There are implications here for intelligence professionals at all levels of business ... and law enforcement and national security, too.  

Let me explain.  Kickstarter is by no means the only crowdfunding site these days.  IndieGoGo and RocketHub are two popular alternatives, but there are a growing number of these sites.  The pace of this growth is likely to increase in 2013 as new laws are set to come into effect that will allow contributors to take small equity interests in start-ups (the current crowdfunding model centers on what typically amounts to pre-sales of a product or service).  Forbes expects revenues generated by crowdfunding sites to double, from 3 to 6 billion USD, in 2013.  Increasingly, this is the way everything from music to games to books to electronics will be funded at start-up.     

Savvy intelligence professionals in the business world should be watching these sites for potential competitors that might emerge from successfully funded projects.  Likewise, given the underfunded nature of most start-ups, it makes equally good sense to see successful crowdfunded projects as a no-cost extension of your own R & D programs - buying out small companies with proven products may well be less expensive than developing them in-house.   Finally, understanding crowdfunding is going to be an increasingly essential element of providing entrepreneurs any meaningful intelligence support.

Law enforcement intelligence professionals should also be taking note.  While there is little evidence of criminal activity in crowdfunding activities to date, with an increase in money, crime is virtually certain to follow.  Fraud of all kinds, money laundering, illegal or unsafe products are all activities which the reputable sites work very hard to avoid but, with the expected growth, criminals will inevitably carve out new niches in the market.

While national security intelligence professionals might not be interested in some artist's new comic book, many of these sites either specialize in or actively promote innovative hi-tech product development and design.  In fact, some academics are even turning to crowdfunding sites to fund their research.  Everything from geo-mapping to verification of information on the internet to pens that can draw in 3D have been funded by crowdfunding.

Crowdfunding is both an innovative and disruptive practice that is set to become much more important in the coming years.  In my opinion, it is worth staying ahead of this particular curve.

Best Professional Development Event For Intel Professionals This Year

That's right, the Institute of Intelligence Studies is sponsoring a professional development conference for its alumni, students and faculty this summer from 8-11 July at Mercyhurst University.

And you can register here!

This is not an alumni "weekend", "retreat" or an "escape".  It is not "homecoming".  This is a premiere opportunity to hone your skills as an intelligence professional at the same place where you all got your start -- Mercyhurst.

We are bringing in top notch speakers (many of which are former classmates!) and are going to offer an extraordinary opportunity to network both inside and outside your chosen field (Quick:  How many Mercyhurst intel alums currently work in the banking industry?  In energy?  In almost any field you can name?  You will be surprised at the number!) and across the generations (We have been doing this for 20 years - I know, I find it hard to believe myself!).

If you need more enticement, the faculty will all be here to bring you up to speed on our current research activities (AKA "What we have learned about intel since you've been gone...") and where we are going in the future.  With over 1000 alumni, 100's of research projects, dozens of full and part time faculty, and the opportunities represented by the new building, these are results you will want to know about.

This isn't just about us, though.  We also want to learn from you, too.  Our program is 20 years old this year and it is time to plot our course for the next 20 years.  What are the problems you are having?  What research do you need done to improve your ability to provide actionable intelligence to the decisionmakers you support?  What skills and abilities do the intelligence analysts of tomorrow need today?  Getting your input at this conference is crucial.

If you are a current student or alum you can register here to attend.  I personally look forward to seeing each and every one of you!

Entry-Level Intel Analysts In Law Enforcement Jobs Report Now Available For Download!

Last fall, we began a quest to capture the hiring prospects over the next 12 months for entry-level intelligence analysts (the kind we produce) in each of the three major sub-disciplines of intel:  National security, law enforcement and business.

Today we are putting out the 2012 Entry-Level  Hiring Projections for Law Enforcement Intelligence report (click on the link to download).

The report was compiled by the same analyst who produced the national security report to rave reviews, Whitney Bergendahl.

This report contains, in addition to Whitney's analysis, the collected wisdom of all the hiring managers and intelligence professionals who took our survey on job prospects.  I would particularly like to thank the International Association of Crime Analysts and the International Association of Law Enforcement Intelligence Analysts for circulating our survey among their members.

Of course, we welcome your feedback (send it directly to me or leave a comment below).   For those of you interested, in December, 2011, we put out the national security report and you can still find the details here.  We look forward to publishing the last installment on the business market for jobs next week some time.

An Internet Murder?

I ran across this remarkable video today and it occurred to me that I might be able to use it to offer a brief but challenging thought experiment:

Given this video, what question would you ask first?

Leave your answer in the comments!

How To Spot An ATM "Skimmer" And Why You Should Care (KrebsOnSecurity)

Skimming is the theft of ATM or credit card information during the course of what appears to be an otherwise legal transaction. ATM skimmers are designed, for example, to acquire the ATM card number and then, through a variety of different devices also acquire the PIN. This allows the thief to collect the data and then use it to get access to the account.

KrebsOnSecurity (via Boing Boing) had a very interesting example of one such skimming device (see picture) with links to pictures of other such devices. A casual search of the internet yielded many, many other examples (including this YouTube video). Lifehacker also linked to a very good PDF by an Australian firm with some detailed info on both the skimmer and the PIN capturing devices.

This type of fraud has been around for some time now and the tricks used by the bad guys continue to get more sophisticated. Despite this, it seems that many people are not aware of the risks. It is worth taking a look at Krebs and the YouTube video simply to be armed with a little bit of info.

MethHunter: Find, Fix And Incarcerate! (DagirCo)

Want to cut your town's meth problem in half next year? You can. Really.

The guys over at Dagirco have finally done it! MethHunter is on the streets and check out these early stats: One of our small, local police departments is using it and they have had over a dozen busts and over a half dozen convictions in the last six months. The best part? The program has cut the time it takes to conduct an investigation in half!

One of local TV stations (WICU) recently did a five part special series on meth in NW Pennsylvania in general and on the MethHunter in particular. The video below is only one part of the five part series. The other four parts can be accessed at the end of this video:



I am particularly proud to say I know the team at DagirCo that put this piece of software together. Most of them, including Mark Blair, the CEO, are Mercyhurst grads and I have had the great good fortune to have many of them in class over the years.

Beyond this, though, I am particularly impressed with the way they tackled the problem. They have essentially created an expert system that looks at pill shopping patterns and "thinks" about them the same way a meth expert might think about them. It automates, as Mark puts it in one of the videos, the "80% of the tedious, time-consuming" work of analyzing purchase records. It even examines the shopping patterns for evidence of denial and deception on the part of the pill shoppers!

It is easy for you to imagine this all getting very technical and difficult to interpret. That is where this program really shines. The DagirCo team has worked particularly hard to make the program user friendly for small town and rural police forces. Their thinking was that large cities often have dedicated meth experts but that small and rural police forces may not have the resources for such a position. This meant that the program had to produce something that was understandable and actionable by a police officer who had received no special training.

It turns out that the straightforward, cop-friendly way the program generates output benefits both small and large police forces. It gives the smaller police force a capability where it had none before and it saves the experts in the larger forces time that can be better spent looking at aspects of the meth problem that the software cannot address.

The DagirCo team is also particulaly proud of the "engine" they developed to drive the MethHunter. They think it can be used in a broad range of applications. Currently, they are thinking about how it might be applied to other problem drugs and even other crimes in general (they are working on an anti-graffitti version right now). Mark, a former marine, also believes a modified version of this software could be useful in analyzing insurgent attack patterns in Afghanistan and elsewhere.

On Being A Spy; Europe, Iran And The Bomb; How Organized Crime Got Hooked On Drugs -- And Other Interesting Stuff (Fora.tv)

Fora.tv is a streaming video service that hosts speeches and presentations by world class experts in a variety of areas. A number of their recent offerings should be particularly interesting to intelligence professionals (Note: The descriptions below come directly from Fora but have been lightly edited for length, etc.):

Stella Remington on Being A Spy
Stella Remington says she's had 4 careers. First as a librarian/archivist, then a diplomat's wife, in MI5, and now as an author. She has written four books, her first a memoir titled Open Secret: the Autobiography of the Former Director-General of MI5. She has published three spy-thriller novels Secret Asset, Illegal Action and Dead Line. She's currently working on a fourth novel.
Date: Fri, 27 Mar 2009
Location: Sydney, Australia, Dymocks Literary Lunch
Program and discussion: http://fora.tv/2009/03/27/Stella_Rimington_on_Being_A_Spy

Under a Mushroom Cloud: Europe, Iran, and the Bomb
Since Iran's illicit nuclear program was exposed to a stunned world in 2002, Tehran has defied the international community and continued to pursue its nuclear goals. What drives this seemingly apocalyptic quest? Are Iran's aims rational or not? Under a Mushroom Cloud analyzes this catastrophic and murky situation, and examines Iran's dual-track approach of accelerating its nuclear activities while weaving itself ever more tightly into the fabric of the European economy.
Date: Wed, 01 Apr 2009
Location: Transatlantic Institute, Brussels, Belgium
Program and discussion: http://fora.tv/2009/04/01/Under_a_Mushroom_Cloud_Europe_Iran_and_the_Bomb

Smack Express: How Organised Crime Got Hooked on Drugs
One of Australia's most successful detectives, Clive Small's book, Smack Express: How Organised Crime got Hooked on Drugs is an insight into drug trafficking and organised crime on Australia's east coast. Written with journalist Tom Gilling, it features an extraordinary range of colourful characters and situations such as one bloke who thought that throwing someone into the boot of a car and driving it to South Australia wasn't kidnapping, because "he never asked to get out of the boot".
Date: Thu, 09 Apr 2009
Location: Gleebooks, Sydney, Australia
Program and discussion: http://fora.tv/2009/04/09/Smack_Express_How_Organised_Crime_Got_Hooked_on_Drugs

Border Patrol: Pakistan and Afghanistan
Eight years after 9/11, the border between Pakistan and Afghanistan is as lawless as ever and Osama bin Laden is still on the loose. Should the U.S. move to secure this region, home to ranks of Taliban and al Qaeda leaders?
Date: Sun, 01 Feb 2009
Location: Foreign Policy Association, New York, NY
Program and discussion: http://fora.tv/2009/02/01/Border_Patrol_Pakistan_and_Afghanistan

Using Craigslist To Fight Crime (StPaulChurch10-21.Wikispaces.com)

The effort to crowdsource a solution to the rash of church burglaries in St. Paul has been making some headway of late. One of the more interesting products to come out of the effort is a whitepaper on how to automate the collection of data using the RSS feeds from Craigslist.

Craigslist is an online classified advertising service that is available for many towns and cities. It is no surprise then that criminals try to use it to fence stolen goods. This is not a new phenomena (criminals use whatever is available and popular to sell stolen property -- pawn shops, flea markets, even, upon occasion, print classifieds). Craigslist cooperates with local law enforcement to keep their trade legit but, given the volume of listings, there is no way they can eliminate all of the traffic.

Being able to quickly scan Craigslist for known stolen items, then, seems to be a valuable tool in the law enforcement analyst's arsenal. Mike Himley, President and CEO of Eagle Intelligence, and a major contributor to the crowdsourcing effort on the St. Paul's churches, has come up with a simple tutorial that allows anyone to set up such a scan. I could easily imagine this as a part of a classroom exercise or as a real tool to help law enforcement identify and recover stolen property.

Even if the St. Paul Church Burglary crowdsourcing initiative does nothing but add these kinds of tools to the analyst's arsenal, it has been more than worth the effort.

Interested In Learning More About Crime Analysis? (Analyst's Corner)

Deborah Osborne, a writer and former crime analyst, sponsors a very interesting podcast over on her blog Analyst's Corner. Her recent interview with Jim Mallard, the crime analysis supervisor for the Arlington, Texas Police Department and winner of the 2008 Technology Award from the International Association of Crime Analysts, is a good overview of crime analysis as a profession, and the role of technology in crime and intelligence analysis. Worth listening to if you are a professional or a student interested in the profession. For me, the most interesting aspect was comparing the problems and techniques of the crime analyst to those in the national security community (with which I am more familiar).

MCIIS Participation In Graffiti Task Force Highlighted (Goerie.com)

In case you missed it in this morning's Erie Times News (doesn't everyone read that?), the paper highlighted the Institute of Intelligence Studies' participation in the anti-graffiti task force here in Erie. The effort is pretty large but we are just doing the analysis part. Interesting work for Professor Dave Grabelski and some of his students!

Annual Wiretap Report Released (USCourts.gov)

The Director of the Administrative Office of the United States Courts released its 2007 Wiretap Report to Congress recently. The statute requires that specific information be reported each year "including the offense(s) under investigation, the location of the intercept, the cost of the surveillance, and the number of arrests, trials, and convictions that directly result from the surveillance."

According to the report: "A total of 2,208 intercepts authorized by federal and state courts were completed in 2007, an increase of 20 percent compared to the number terminated in 2006. The number of applications for orders by federal authorities fell less than 1 percent to 457. The number of applications reported by state prosecuting officials grew 27 percent to 1,751, with 24 states providing reports, 1 more than in 2006."

However, there is a catch. The Feds did not include data "involving sensitive and/or sealed matters". That means the report had to include this little gem: "Statistics indicate that if all intercepts undertaken for federal investigations in 2007 were reported, the 2007 Wiretap Report would not reflect any decrease in the use of court-approved electronic surveillance by the agencies." Given that Federal wiretaps reported are below the ten year average, it is virtually certain that the actual number of wiretaps is higher, though probably not hugely so.

The report goes on to say: "Installed wiretaps were in operation an average of 44 days per wiretap in 2007, compared to 40 days in 2006." This came at an average cost of over $48,000 per wiretap! The most common specific location for the wiretap was a “portable device, carried by/on individual” by a very large margin -- 94% of wiretaps were on cell phones, etc.

81% of all applications for wiretaps stated that drug offenses were the most serious crime under consideration. There is not even a category for terrorism on the chart provided. It is unclear if this is because terrorism is included in the "other" category (which, at the federal and state level combined, accounted for only 30 wiretaps) or because these are "sensitive and/or sealed matters".

Arrests vs. conviction rate told the most interesting story to me, however. The graph below (built by me using the data from Table 9 of the report) shows the percent of convictions in cases where wiretaps were used versus the total number of wiretaps authorized per year (I have normalized the number of wiretaps by dividing the total by 100 so that both lines could easily fit on the same graph. To get the actual number of taps, you need to multiply the raw number by 100). I have also added a trendline to both sets of data. It appears that the number of arrests is staying more or less the same but that the number of convictions is dropping. This might just be a statistical artifact (arrests are often made months or years after the intercept is made) but it is still a trend worth noticing.

Law Enforcement Intel: Materials Theft Background And Resources (CSO, ISRI via LII)

I don't write much about law enforcement intel here. We have three experts on staff at Mercyhurst and others (such as Deborah Osborne over at Analyst's Corner and David Jimenez at Intelligence Is The Future) cover this topic with much more authority than I do.

I ran across these two articles on materials theft recently, however, in the Librarian's Internet Index (an excellent source of reliable open source material) and, since I had not seen them elsewhere, I thought I would write a quick post about them.

Red Gold Rush: The Copper Theft Epidemic is a good introductory article concerning the process of metal theft including some good intelligence indicators like this one: "What he had found was a regular burn site for metal thieves. In fact, it is the perfect burn site, with a concrete surface to burn on and available running water to control and put out fires. Also, the site is surrounded by metal to steal. A nearby communications tower had already been looted so much that, at one point, 911 service was knocked out. Plus, there is a scrap yard nearby where the stolen metal can be sold."

The Institute of Scrap Recycling Industries also seems to provide a number of good resources about how to identify and prevent metal theft including a document on Recommended Practices and Procedures for Minimizing the Risks of Purchasing Stolen Scrap Materials.

Mercyhurst College Students Help Bust "Oldest" Meth Ring (Goerie.com)

Some of our students, working as interns for the local Methamphetamine Task Force provided the analysis that helped take down a methamphetamine "super lab" yesterday. The news story is here and the College's press release with a bit more info on the student's role is here. The video below only briefly mentions Mercyhurst but gives some good background to the bust. Congratulations to the students for their very fine analysis! (That's what I call an interesting internship...).