Forensic Virtual Models- Bridging the Gap to Virtual Reality

In this VizWorld Feature, Eugene Liscio of AI2-3D returns with an article discussing the power of Forensic Virtual Models in the Courtroom.

When virtual reality became popularized in the late 1980s and 1990s, it wasn’t long before people started to dream of fascinating concepts and uses for virtual environments in education, simulation, training and entertainment. However, at the time the progression of technology fell short in terms of processing power, responsive tactic input devices and the fundamental programming and logical techniques required to output a realistic looking environment.

Today, there have been large improvements in processing power as well as the techniques used to display cutting edge 3D graphics in real-time. By simply looking at the gaming, medical and engineering design markets, one will quickly realize that the developments in computer graphics in the last 20 years has been astonishing. Yet, even with the advancements in processing power and logical programming techniques, the use of virtual reality in the forensics and litigation fields has seen minimal use.

Read more after the break.

There may be many reasons for this, such as the fact that the legal industry is slow to make large sweeping changes when it comes to new technologies. There needs to be due process and a period where technologies are explored for their scientific foundation and the risk of unfair prejudice is always in question. Also, cost is often a consideration and the necessary hardware required to “suit up” a jury of 12 with headsets, data suits and other devices can be cost prohibitive. In all, the central reasons why virtual reality has not seen widespread use in our legal systems centers around a technology that has not been accepted industry wide for use by many in the forensic and legal system.

In the meantime, there have been many other technologies that have arisen along the lines of virtual reality that are not so overwhelming but have accomplished much of the same things but on a lesser immersive scale. Many people are familiar with websites that now host virtual models of vehicles, clothing, artworks and other objects of interest as rotatable 3D models. These interactive models are an excellent way to communicate a product without actually holding a physical object in your hand or driving to the mall to get up close for an inspection.

Also, 3D objects no longer need to be static models that are simply rotated around for viewing. They can now be animated by the simple click of a text button or they can even provide useful measurement information when clicked in certain parts of a model. In the field of forensics, these models have come to be known as Forensic Virtual Models.

Forensic Virtual Models (FVMs) are the next evolution of forensic animations in litigation that are helping to bridge the gap between animation and virtual reality. They can best be explained as the migration of 3D content (i.e. the data used to create a forensic model or animation) into an interactive application with a real-time graphics visualization engine. Unlike any regular videos, film or forensic animations which are passive (i.e. the user sees only what the animator wants to show), FVM’s are viewed in real-time and the user can interact with the animation by moving the camera’s view in the scene as the animation is playing. The visualization engine may often allow for smart links and object interactivity so that when the user gets close to or interacts with an object, it can react in a pre-specified manner.

Traditionally, forensic animations have been used in forensics and litigation to present the findings of a case to courts and jurors for better interpretation of the facts and theories of a particular event. However, FVMs are a rather benign departure from traditional 3D animations and extend their use into the world of virtual objects and environments without the user becoming completely immersed in an “unfamiliar reality”. The transition from a forensic animation to a FVM is quite simply a smaller step that is easier to adopt and accept in a courtroom scenario.

There are in fact, many parallels between forensic animations and 3D FVM’s. They both have a very similar initial development process, they are both mainly driven by technology from the computer gaming industries and they are governed by similar rules of admissibility. However, the main difference between the two is that a forensic animation is a passive visual experience whereas the FVM’s offer the user a dynamic and interactive experience which is on parallel with modern day video games.

Unfortunately for some, any mention of the gaming industry tends to provoke negative thoughts for use of the technology in Forensics. One would believe that it is the irony of the matter that the same software used to create some incredibly creative and extraordinary games can also be applied to an industry where the creed is “adherence to the facts”. However, it is the gaming industry that is responsible for driving some of the most stunning advancements in 3D technology used in the medical, military and visualization industries. During the late 1990’s when forensic animations began to grow in use, there were similar arguments against this tool being used in court due to the apparent prejudicial nature of this form of demonstrative evidence. Today, most of these issues are better understood and have been overcome by a better educated industry. The use of forensic animations is widespread across the industry and is frequently used for analysis and in courtroom presentations by accident reconstructionists, medical experts and crime scene investigators.

FVMs do not require specialized hardware or software to operate (although some minimum requirements should be met). FVMs are viewed on a regular computer screen and are not intended to be fully immersive environments, but rather an interactive platform to inspect and gain knowledge of a particular scene, event or model. There are no special devices required to interact with the virtual models, although this could be incorporated much the same as a game controller is used to operate a person’s movements or a vehicle in a video game. Typically though, the main controller is the computer mouse and keyboard. Therefore, this makes the FVMs easily viewed and operated by a multitude of people simply by sitting at their computer screens and launching the application or viewer.

The benefits of a FVMs are that forensic experts no longer need to have a physical object in their hands to present to a jury. Often, the objects and evidence in a crime or accident case can be potentially dangerous in the courtroom and are better left in storage or elsewhere. At other times, the evidence is simply impossible or impractical to transport such as pieces of an aircraft recovered in a tragic accident or where there is blood spatter and other evidence which cannot be transported without being tainted.

Figure 1. Section of a concrete wall and floor where bloodstains (highlighted) can be recreated as virtual model so that they can be analyzed and viewed from different angles many years after an incident has occurred.

Another example is in the case of a medical examinations where the issue is actually in several images of an X-Ray, CT or MRI scan. The images captured are entered as evidence, but can also be used to create 3D interactive models. In the case of CT and MRI images, it is possible to directly convert the images (when in the proper format) into a 3D model for inspection of a particular feature. This type of evidence would be impossible by any other means. For X-ray or single images, they can be used to scale and locate anatomical features. All this can then be packaged as a FVM and presented to a jury who will find looking at an anatomical 3D model much more familiar than looking at a number of CT and MRI images that will inherently be somewhat abstract and unfamiliar.

Figure 2. Screen capture of a Forensic Virtual Model created based on a CT scans. The FVM can be used for inspection, analysis and courtroom presentations.

For crime scene recreation, FVMs provide a sense of position, scale and perspective relationships that aid the investigator and jury in determining the point of view and sequence of events. Complex scenarios that have multiple events happening at the same time can all benefit from watching and event take place from different perspectives and locations throughout a scene.

Figure 3. Screen capture of a Forensic Virtual Model for a shooting recreation.

There are of course limitations to the nature of what is being presented through FVMs in that there are typically no “artificial intelligence” systems in the FVM itself. The objects and environment are built based on as much known data as possible. Also, the virtual environments are not simulated in any way. This means that there are no physics engines, weather systems or dynamic lighting running or controlling the scene environment. All the movements of objects and placements of models are under the full control of the animator. This is quite acceptable as it would be highly undesirable to have varying or unexpected scenarios while presenting a case in a courtroom. Therefore, only what is consciously placed into the virtual environment by the animator can be shown to the viewers. The scene, environment and movements of objects is repeatable and consistent.

There will no doubt be a place in the future for fully immersive virtual reality in litigation. With the advent of new 3D displays and optic input devices that remove the need for data suits and special head gear, the likelihood is that virtual reality is not so far off in the future as some might believe. Acceptance of virtual reality is also dependant on the successful implementation and use of intermediate technologies. For the moment, FVMs are assisting in bridging the gap between animations and virtual reality.

Eugene Liscio, P.Eng and owner of  AI2-3D,  specializes in accident reconstructions and forensic animations to be used by police and lawyers in prosecuting cases.  They have several example animations on their site and Youtube.

Eugene was kind enough to provide to VizWorld and its readers a sample Forensic Virtual Model, the one shown above in Figure 3, for you to download and try on your own machine.  You can download it here (17MB).

Note: The FVM requires Windows and DirectX9 with new drivers.  I’ve been unable to get it to work on XP64, but been successful on Vista.