Tattoos weren't visible on security camera footage either, so when Oregon police officers got a tip that the culprit may be a man named Tyrone Allen, who does have facial tattoos, they edited his tattoos out of his mugshot before showing it to eyewitnesses in a photo lineup.
Tyrone Allen (top left) has tattoos on his face, but police edited them out of the photo they used in an eyewitness lineup (top right) to make him look more like a bank robber (bottom right).
But according to his attorney, the fact that police manipulated his photo to make him look more like the perpetrator makes the identification evidence unacceptable.
"By presenting an unnecessarily suggestive fake picture of Mr. Allen to the bank tellers, investigators created a substantial likelihood of misidentification," assistant federal public defender Mark Ahlemeyer argued in a recent court filing.
The case highlights the controversial but technically standard practice of police editing lineup photographs, which are already problematic because of the fallible memories of traumatized victims. According to the Innocence Project, mistaken eyewitness identifications have contributed to more than 70% of wrongful convictions later overturned by DNA evidence.
Mat dos Santos, legal director at the Oregon American Civil Liberties Union, said the fact that police didn't acknowledge the edits until Ahlemeyer brought it up raises questions about how prevalent it really is.
"We the public should be asking for an investigation into this practice," dos Santos said, "so that courts can determine whether people have been wrongfully convicted or suffered procedural problems sufficient to grant them a new trial."
Prosecutors in Allen's case have said that at the time of the edits "the perpetrator was potentially still at large," suggesting Photoshop would give them a way to quickly identify the culprit. They have also argued that the editing might have actually made the lineup more fair because leaving Allen's face tattoos in "would have singled him out."
Ahlemeyer, on the other hand, has pointed out that "the tellers did not actually identify a picture of Mr. Allen," in his motion to suppress the evidence.
"Instead, they selected an image that had been altered to make Mr. Allen appear different than he actually does," he wrote.
Multiple U.S. Supreme Court rulings have sought to protect due process by barring identification evidence procured by unnecessarily suggestive law enforcement procedures. According to Melissa Colloff, a forensic psychology professor at the University of Birmingham in England who's published research on photo lineups, a guiding principle is that suspects should not stand out.
To ensure that every lineup member is a plausible alternative to the suspect, she told Law360, police departments have a few options. A 2003 U.S. Department of Justice manual on eyewitness evidence recommends replicating a suspect's distinctive features onto the other "filler" photos in a lineup.
"If the suspect has a unique feature not described by the witness, you should not alter the suspect's appearance," the manual says. "Rather you should select fillers that have a similar, but not identical, feature or enhance the fillers with a similar feature."
In Allen's case, that would have meant editing the other photos to have facial tattoos too. But in court filings, the government said that would have been problematic "because the photos would have been inconsistent with the descriptions provided by the victim tellers, none of whom reported seeing facial tattoos."
According to a 2004 survey of police practices, Colloff noted that 77% reported trying to find fillers who had similar features. Only 23% reported using replication — and 18% reported using concealment, as the Portland police did.
She said that, on first glance, it seems like concealing Allen's facial tattoos helped make the lineup more "perceptually fair" because the suspect did not look noticeably distinct from the other lineup members.
"But the concern here is that, in doing so, officers potentially made the suspect look more like the witness's memory of the culprit (i.e., without tattoos)," she wrote in an email.
The other concern, dos Santos says, is that police didn't mention the altered photographs to witnesses or Allen's attorney.
It wasn't until discovery proceedings that Ahlemeyer noticed how Allen's booking photo, taken in connection with unrelated warrants, had been altered to conceal his facial ink.
"When I read this, I thought to myself, 'wow this an outrageous practice,'" dos Santos said. "Portland police should reveal to us every time that they have made this call to alter defendant photographs to more closely match eyewitness testimony."
He added that the failure to disclose the alterations before discovery could amount to a Brady violation, which occurs when prosecutors are aware of exculpatory evidence but do not share it with the defense.
A hearing on Allen's motion to block the government from using the lineup evidence was held on Aug. 14; the judge has not yet ruled.
Prosecutors in the case did not respond to a request for comment, but their court filings have noted investigators saw enough similarities between Allen and the man shown by bank security cameras to include him in the array, "knowing that tattoos are easily concealed with make-up."
They also cited U.S. v. Ellis, a 2015 case in which police altered the images of six men in a lineup to make them all appear to be wearing a hoodie similar to one worn during the crime in question. A judge signed off on the alterations, finding that the changes made identification "more reliable, not suggestive."
In response, Ahlemeyer said that case would be only applicable to Allen's case if police had edited the photo array to show each person wearing a baseball hat.
"In contrast to Mr. Allen's case, at no point did the investigators in Ellis change any aspect or feature of the suspect's face," he said in a filing Aug. 19.
"The question in this case," he said, "is whether the police can alter a photo on their own accord to remove distinctive facial characteristics that conflict with objective video evidence and eyewitness reports ... the answer should be no."
Have a story idea for Access to Justice? Reach us at firstname.lastname@example.org.