Times Square, Bombs and Big Crowds

car bombNew York Police Department/European Pressphoto Agency Image from a surveillance camera showing the Nissan Pathfinder in Times Square on May 1. View more photos.

Updated, 8:55 a.m. | Jeffrey Rosen, a law professor and author of “The Naked Crowd,” joins the discussion.

Updated, May 4, 7:50 a.m. | Federal agents arrested a Connecticut man shortly before midnight on Monday for driving a car packed with explosives into Times Square on Saturday evening. The suspect, Faisal Shahzad, a naturalized United States citizen from Pakistan, was taken into custody at Kennedy Airport on board a flight to Dubai.


The police and F.B.I. investigators have tracked down the person they believe to be the owner of the Nissan Pathfinder that contained the makings of a crude car bomb discovered in Times Square on Saturday evening. He is a naturalized U.S. citizen from Pakistan who recently returned from a trip there. In identifying the man and whether he or others might have been involved in the attempted bombing, the police reviewed surveillance footage of pedestrians in the area.

Do security cameras work? How can we prevent terrorist attacks in crowded areas?


Don’t Panic, Get Used to It

Richard Clarke

Richard A. Clarke was national coordinator for security and counterterrorism for Presidents George H.W. Bush and Bill Clinton. He is the author of five books, including “Cyber War: The Next Threat to National Security and What to Do About It.”

Like the Fort Hood attack, the Times Square attempted bombing may be — according to what I’m hearing from senior White House officials — another example of sympathizers of a foreign terrorism movement living in the United States, becoming persuaded of the need to act and then doing so without training, using readily available materials and acting on general guidance obtained over the Internet from abroad.

Homegrown plots are highly likely, and some will succeed.

Such “homegrown” attacks will not typically pose the same level of threat that can come from a large scale, coordinated mission carried out by terrorists trained together in foreign camps. These semi-spontaneous attacks are, however, much more difficult to detect in advance. There are little or no cell structures to infiltrate, few warnings and minimal signs of preparation.

Despite the presence of the largest and best urban police force in the nation, someone was able to drive a potential car bomb into the middle of Times Square, park it with hazard links blinking, jump out and escape.

Read more…


Put More Cameras on the Streets

Steven Simon

Steven Simon is an adjunct senior fellow in Middle Eastern Studies at the Council on Foreign Relations and the co-author of “The Age of Sacred Terror” and “The Next Attack.”

Video surveillance would not have stopped the Times Square attack. Does this mean that it would be useless? Not necessarily.

The challenge: exploiting this visual information while protecting the privacy of citizens.

Swift and accurate analysis of video surveillance information might prevent the next attack, even if it is powerless to stop the last one. Imagery can be used to assist in the identification and location of individuals at the scene of the crime.

It can also be used to track the progress of the bomb-laden vehicle from the its point of origin, or the point at which the truck was weaponized, to the place the terrorists have targeted. In combination with physical evidence acquired from the vehicle — fingerprints, hair, cloth fibers, soil, trash, forgotten personal items or a host of other bits of evidence — video surveillance can lead to the arrest of the bombers and to the unraveling of cells or networks and, if the attackers are foreign, the ratlines they exploited to enter the country.

Read more…


More People, More Training

Paul Ekman

Paul Ekman, professor emeritus at University of California at San Francisco, is the head of a small training company the Paul Ekman Group.

Training the New York Police Department Counter Terrorism Division this past winter about how, why and when observing demeanor can suggest a potential problem, an impending attack or lying about a serious matter, left me both worried and reassured.

Surveillance cameras are not the answer to prevent attacks.

Worried because New York City is a prime target and spotting an attacker is like looking for a needle in a haystack. It takes only one missed attacker to result in crippling damage.

Reassured because the department’s Counter Terrorism Division is the most sophisticated, dedicated, hard-working, ingenious organization I have encountered in law enforcement. You could not ask for more, but it may not be enough.

Read more…


Measurements by Camera

Michael Black

Michael J. Black is a professor of computer science at Brown University. His research in computer vision focuses on video motion estimation and human shape analysis.

A surveillance camera catches a suspect walking down the street. Unfortunately the quality is not good, the suspect is far away, their face is not visible, and the lighting is bad. What we want is evidence. What we have is a low quality video.

Image analysis won’t be able to pick out a known terrorist in a crowd, but it could rule out a possible suspect.

Forensic video analysis addresses the extraction of evidence from video. Each frame in the video is simply a collection of numbers describing the color of every pixel. To be useful to law enforcement, that collection of numbers has to be converted into evidence. How tall is the suspect? How much do they weigh? In a sense, we need to turn the camera into a measurement device.

Consider the problem of estimating how tall the suspect is. You could count how many pixels high they are but this would not tell you their height in inches.

Read more…


What Electronic Eyes Miss

Noah Shachtman

Noah Shachtman is a contributing editor at Wired magazine and editor of its national security blog, Danger Room.

On March 6, 2008, a man in a gray hooded sweatshirt set off a small bomb in Times Square. There were hundreds of surveillance cameras around waiting to catch him on tape, but the bomber was not deterred.

Surveillance software works well in controlled environments like casinos, but is less reliable in the chaos of city streets.

This past weekend, the scenario repeated. A second would-be terrorist tried to strike Times Square, this time with a car bomb. Once again, the cameras did little to dissuade the attacker.

So it’s more than a little ironic that some city officials are now suggesting that the answer to protecting Times Square is … more cameras. New York is using a $24 million federal grant to assemble a network of public and private security cameras and license plate readers.

A single command center for this Midtown Manhattan Security Initiative will gather together all the feeds, and use advanced algorithms to catch suspicious cars – and suspicious behavior – before a terrorist detonates any explosives.

Read more…


A Long Way From Human Intelligence

Michael Tarr

Michael J. Tarr is the George A. and Helen Dunham Cowan Professor of Cognitive Neuroscience and the co-director of the Center for the Neural Basis of Cognition at Carnegie Mellon University.

Security cameras are only as good as the brains behind them. Even if the video is pristine, without some prior knowledge of the suspect’s identity all you end up with is a nice wanted poster.

Machine vision has made progress, but there are limits on how much information can be extracted from poor data.

That being said, typical surveillance video is noisy — faces are obscured by disguises, other objects and people, and the low image quality (as demonstrated by the Times Square video) of most surveillance cameras. Machine vision has made progress cleaning up such images and video, but even still there are limits on how much information can be extracted from poor data.

Indeed, this is one of the many areas where human intelligence still has a huge advantage over machines. As individuals we are incredibly good at face recognition — experts if you will — able to identify more than 10,000 people across variations in lighting, viewpoint, and facial configuration.

Read more…


Focus on the Threat

Bruce Schneier

Bruce Schneier is a security technologist and author of several books on computer security, including “Beyond Fear: Thinking Sensibly About Security in an Uncertain World.”

In the wake of Saturday’s failed Times Square car bombing, it’s natural to ask how we can prevent this sort of thing from happening again. The answer is stop focusing on the specifics of what actually happened, and instead think about the threat in general.

Our approach needs to be flexible and adaptive, just like the bad guys.

Think about the security measures commonly proposed. Cameras won’t help. They don’t prevent terrorist attacks, and their forensic value after the fact is minimal. In the Times Square case, surely there’s enough other evidence — the car’s identification number, the auto body shop the stolen license plates came from, the name of the fertilizer store — to identify the guy. We will almost certainly not need the camera footage. The images released so far, like the images in so many other terrorist attacks, may make for exciting television, but their value to law enforcement officers is limited.

Check points won’t help, either. You can’t check everybody and everything. There are too many people to check, and too many train stations, buses, theaters, department stores and other places where people congregate. Patrolling guards, bomb-sniffing dogs, chemical and biological weapons detectors: they all suffer from similar problems. In general, focusing on specific tactics or defending specific targets doesn’t make sense. They’re inflexible; possibly effective if you guess the plot correctly, but completely ineffective if you don’t. At best, the countermeasures just force the terrorists to make minor changes in their tactic and target.

Read more…


A Future of 24/7 Surveillance

Jeffrey Rosen

Jeffrey Rosen, a law professor at George Washington University and the legal affairs editor of the New Republic, is the author of “The Naked Crowd: Reclaiming Security and Freedom in an Anxious Age.”

After every thwarted terrorist attack, there are predictable calls for more surveillance cameras. Before the failed Times Square bombing attempt, the N.Y.P.D. got a $24 million Homeland Security grant to install a security network in Midtown with surveillance cameras, chemical sensors, and license plate readers. Despite evidence that such a system wouldn’t have prevented last Saturday’s attack, many will demand an even more elaborate surveillance network as a result. But all the best empirical research suggests that it will be a waste of money.

Consider how Google, Facebook and other Web services could be harnessed to surveillance cameras to monitor everyone’s movements.

Here’s what we know about surveillance cameras. They’re more or less useless in deterring terrorism before it occurs. The best peer reviewed studies in Britain and America find no connection between the proliferation of cameras and the deterrence of serious crime or terrorism. They’re also not useful in preventing attacks in progress: the Times Square bombing was detected by alert street vendors who saw the smoke and called the police.

Cameras sometimes play a supporting role in identifying the perpetrators after an attack has occurred. But in all the major terrorist attacks since 9/11, including the London bombings, the perpetrators would have been identified without the cameras. In the Times Square case, there was so much forensic evidence at the crime scene that the police were able to identify the former owner of the Pathfinder through the vehicle identification number, leading to the arrest of the suspect on Monday night without necessarily relying on
the footage from more than 80 surveillance cameras.

Read more…