A video forensics expert would evaluate and enhance the surveillance video.

Hollywood Not, But How Real Forensics Solves Crimes

For over a decade now, countless shows have been entertaining us with the tools and gadgets behind forensics: the scientific side of crime solving. These instruments and techniques are highlighted in shows such as CSI, Law & Order, NCIS, Bones, Without A Trace, and many others. There is little doubt that shows such as these have not only entertained us, but have even changed the way the public perceives crime solving and forensics analysis in particular; but are the public’s perceptions accurate or even fair? It may be important to explore how real forensic science solves crimes.

The Crime

A bank has just been robbed. The robbers meticulously planned and executed their brazen daytime heist. When it was all over, they sped away in their shiny black Dodge Charger with what is easily a couple million in untraceable cash, stuffed into several large duffle bags. What the crooks did not count on, however, was the simple deli across the street having a surveillance camera positioned in such a way that it catches the Charger as the robbers make their getaway.

Enter the Crime Scene Investigators! Frustrated at a lack of evidence in the bank, they start looking for some outside the bank and notice the deli’s surveillance camera. A quick analysis of the footage produces a grainy pixelated clip of several figures exiting the bank and hopping into a vehicle before speeding away. This is no problem for these crime scene investigators. A few clicks and commands into their super forensics computer and what was once a barely recognizable video becomes a clear hi-definition video! They zoom in on the bad guys faces and wa-la! All that needs to be done is plug that into some facial recognition software and they’ll have the identities of the robbers in seconds.

Unfortunately, this is a highly romanticized and exaggerated narrative of how modern day forensics experts analyze and enhance video evidence in real life situations. Let’s take another look at the video footage and see what we can really do.

Video Forensics / Still Image Forensics

First of all, it’s important to define pixels. Pixels are the smallest unit of visible data which comprise an image. The image could be a still picture or a video, but whatever it is can be broken down to these pixels. The higher a pixel count the original image or video has, the higher the resolution. The higher the resolution, the clearer the image or video is. Forensics experts can only work with the number of pixels that a video has at the time it was recorded. Not only that, but as video surveillance evidence is transferred from a Digital Video Recorder (DVR) to a Flash Drive, the data is often compressed which can also result in a decrease in resolution (e.g. less pixels).

So how does this affect the video of our bank robbers? If the video was grainy and pixelated because it was recorded in a low resolution, no amount of forensic manipulation can un-pixilate it. However, modern forensics can utilize cutting edge software and techniques to get lots of vital and useful data from our video.

Firstly, we may start by focusing on the car, and particularly on the license plate. Video forensics experts can sometimes enhance video characteristics around the license in such a way to discern the characters on the plate. Next we can enhance the video around the robbers themselves. Sometimes it can be cleaned up enough to get a clean facial image. Other times we can discern unique facial characteristics: facial hair, bone structure, skin tone, hair color and/or length, freckles, etc. We can also use video enhancement techniques to determine the height of a suspect or the speed of their car. All this information could be crucial in both in tracking down suspects and as evidence at trial.

In some cases, we’re able to utilize a cutting edge technique known as Frame Averaging. Frame averaging allows a forensic video expert to combine multiple frames and multiple details from a surveillance recording and combine them into a single, high resolution image. Using frame averaging, we may be able to get a much more clear, usable image of a robbers face.

Synchronization is the combining of multiple sources of evidence to clarify an event. In other words, we take multiple videos from multiple cameras and synchronize the footage into one cohegent video clip that best shows the events as they unfolded. We are also able to use all available video evidence to map out the sequence of events. This makes it easier to understand the events that took place and the order in which they occurred; this is extremely helpful evidence in court.

What if the video clip shows someone or a car moving very fast, too fast to understand the movement of a body or the interaction of cars in an accident? We can use slow motion, zooming into footage, and repetitive viewing to help understand the details of a case.

Audio Forensics

What about audio evidence? We can clarify audio by removing background noise or by lowering a very loud foreground sound or voice. Sometimes we can isolate sounds, for instance a conversation in the background.

There are also technical aspects of video evidence that often need to be clarified to juries. This may include details such as verifying the date and time stamp on a video, determining the video frame rate, and teaching a jury how wide angle lenses or telephoto lenses may alter a viewer’s perception of what they are watching on the TV monitor.

And finally, what about Facial Recognition software? Facial recognition is one area of forensic science that has seen tremendous improvement over the past decade. The FBI’s Next Generation Identification (NGI) system is cutting edge in the field of facial recognition. This technology is most effective when utilizing images taken under ideal circumstances, such as ID photos or “mug shots”. Potentially blurry images of a moving suspect taken from a distant camera with low resolution may be very difficult to enhance in order to extract identification. Nonetheless, it is certainly foreseeable in the near future that this type of technology will be improved and its uses increased throughout the world of forensic analysis.

Real life is not always as exciting as Hollywood would have us believe. Sometimes, it’s much harder to get the same results that TV and the movies get. But forensics has come a long way. We are able to uncover things today that was never possible not that long ago. While the field of forensics and crime scene analysis continues to advance and change, we expect to see many more astonishing and helpful techniques that will continue to better both the forensic and scientific communities.