Sunday, December 7, 2003
Computer-Generated Image Evidence May Be Here To Stay
By Scott Sandlin
Journal Staff Writer
The Chinese proverb that "a picture is worth a thousand words" got a major workout in a Taos murder trial in which prosecutors used computer graphics to undermine a claim of self defense.
The case against Manuel Tom Tollardo, who shot and killed two men he said were going to attack him, has run the full gamut of the state courts with the Supreme Court's decision not to review an appellate opinion.
His conviction stands.
And the courts have spoken: Computer-generated images as evidence are here to stay. With some caveats.
"In this case, we are concerned with the techniques used to generate computer images. ... The critical inquiry is whether the method used to generate the images is a valid application of the principles of computer technology," Judge Michael Bustamante wrote for the New Mexico Court of Appeals.
The Court of Appeals said the district court was correct in admitting, over defense objections, a series of computer-generated images. The graphics were prepared for prosecutors using "reverse technology" to determine the likely trajectory of gunshots.
It was an issue the New Mexico court hadn't previously addressed.
But courts across the nation increasingly face questions about the admissibility of computer graphics.
Most are trial courts.
Only a few appellate courts have grappled with the issue, said Ed Imwinkelried, a national expert on scientific evidence and law professor at the University of California/ Davis.
It takes time for cases to percolate up to the higher courts, he said. And the more controversial an issue, the more likely it is to be litigated.
"The truth of the matter is, there is an emerging consensus with respect to computer-generated evidence," he said. And the emerging consensus is that it should be admissible.
Is seeing believing?
The initial controversy, which has now subsided, was whether the whiz-bang gimmickry of such evidence would overwhelm a jury.
"My assumption has always been that if we have jurors drawn from a citizenry which views things like 'Terminator 3,' which sees not only things that didn't happen but things that can't happen on the screen these aren't people who are going to automatically buy into anything they see. They will be skeptical enough to second-guess it," he said.
Computer-generated evidence boils down to two basic categories, he said:
Demonstrative evidence illustrates a point in the same way a hand drawing would do and typically doesn't go into the jury room when deliberations begin.
Substantive evidence, more often challenged in this context, is derived from sophisticated computer programs that take in data and spit out images based on equations programmed into the software.
That was precisely the question in Tollardo.
Tollardo, 20 years old at the time of the incident in July 2000, was awakened at his brother-in-law's home by his sister banging on the door, crying, bruised, battered and missing her house keys. She told her brother she had been raped and beaten by Robert Miera, one of her coworkers, and Jeremy Trujillo, a friend of his who claimed to have ties to a Los Angeles gang.
Tollardo, accompanied by his sister and a friend, went to Miera's neighborhood to question him. During the subsequent shouting match in the darkened mobile home park, shots were fired. Both Miera and Trujillo died.
The defense said Tollardo and his sister had been threatened, that a knife was found next to Miera's body and that a sheathed knife was found on the driver's seat of Trujillo's car.
District Attorney Donald Gallegos, who had seen a Discovery Channel report about computer-generated graphics, wanted that kind of analysis for the Tollardo case.
"I needed it because we were having problems with where the wounds were," Gallegos said. "His defense of course was self defense. So what I wanted to do was line up the trajectories to get an idea of how the victims were standing (and) what posture they were in. He said one of victims was reaching for a knife, so we wanted to try and dispel that theory."
If it had happened the way Tollardo said, the bullet holes would have been different, Gallegos said.
"As it was, it looks like that particular victim was facing towards him in a crouched position instead of reaching inside the vehicle."
Gallegos enlisted the aid of the FBI, through a rural assistance grant, to produce the $50,000 graphic he could not have afforded otherwise.
Carl Adrian, a visual information specialist at the FBI lab, used information from police officers and other investigators to determine where the shooter was standing.
His video shows a series of moving, computer-generated figures with dotted lines indicating the bullet trajectories and where the shooter would have been standing.
The defense said the computer graphics didn't meet the state's test for the validity of scientific evidence and argued they should be excluded. The defense also objected on grounds of accuracy and fairness, noting, for instance, that in the video, the shooter is clothed while the victim is naked.
Since Tollardo never denied taking the shots but insisted they were taken only in self-defense, the video offered nothing useful for the jury and in fact unfairly prejudiced the case against him, the defense argued.
"That left only Tom's intent at issue," appellate defender Vicki Zelle wrote. At best, the computer graphics' value as proof was limited "given that the video painted a hypothetical of 'what might have happened' at certain 'freeze-framed' moments."
District Judge Peggy Nelson permitted the video to be shown after a hearing at which Adrian testified and was cross-examined about both his methods and the information he relied upon to generate the graphics.
Prosecutors argued on appeal that the exhibit was a demonstrative aid and they shouldn't be required to meet the test for scientific reliability.
Both Zelle and Bustamante took note of the power of visual images in their discussion of the issue. And both used the picture-worth-a-thousand-words proverb.
"Studies show that jurors retain more information from visual presentations" or a visual-verbal combination than from words alone, Bustamante wrote for the court.
The trial court, he said, "correctly required the proponent of the images to establish the validity of the computer programs used to generate the images."
Allegra Carpenter, an Albuquerque trial lawyer who lectures nationally about using computer-generated graphics at trial, hailed the opinion as providing a valuable tool.
"It was wonderful to finally have a written opinion that discussed some of the nuances of this computer evidence, because the rules have always been a little vague on the difference between substantive and demonstrative evidence," she said.
For example, a Power Point demonstration with little boxes moving around a screen to show how an auto accident occurred is just a visual aid. "It's just a fancy diagram," she said.
It's very different, she said, from a graphic recreation of the accident generated by an engineer who has measured skid marks and the size and weight and resiliency of the vehicle to suggest, for example, the speed at which the vehicles were traveling and their positions.
Carpenter has used computer graphics in criminal cases, including one murder defense, as well as to explain a theory of liability in civil cases, like the complex mine smelting operation where a miner was burned to death.
Imwinkelried, co-author of the text "Scientific Evidence," is also enthusiastic about the emerging rules on computer evidence.
Under the old rules of evidence, expert scientific testimony was admissible only if the principles on which it was based enjoyed "general acceptance" in the scientific community. That, in Imwinkelried's view, permitted lawyers to "count noses" to determine validity.
That changed in the wake of a 1993 U.S. Supreme Court decision dealing with such evidence, Daubert v. Merrell Dow Pharmaceuticals. Now, expert technical, scientific or other testimony is permitted if it aids understanding of an issue and is grounded in scientific method. The new, more flexible standard means a judge determines the validity of individual scientific or technical evidence at a pre-trial hearing.
"Across the board prosecutors, criminal defense attorneys, civil attorneys, judges (legal professionals) have had to learn more scientific methodology. I think it's all for the good," he said.
"Is it hard? Yes. But for decades we were asking the wrong question.
"We were basically asking, if we went to a scientific convention and asked for a show of hands, how many hands would go up? Well, that's no way to figure out whether you have reliable testimony."
Now, he said, attorneys need to roll up their sleeves and ask, What's the size of your database? What was (its) composition? What were your test conditions? What were your validity rates?
"Basically," Imwinkelried said, "we've been running and hiding from science for decades. And now we're finally having to confront it.
"As jolting as that reality is, the justice system will produce more justice now that we've been forced to make this change."