Remote identity validation rally enable people to assert their identity online, without going to a physical location. It is important to understand the performance of these systems and the degree to which they may reduce fraud while maintaining access to services.
The Remote Identity Validation Rally (RIVR) challenged industry to demonstrate secure and accurate remote identity validation technologies that can combat identity fraud when users apply for government services, open bank accounts, or verify social media accounts.
The RIVR was divided into three tracks, for three different steps in the remote identity validation process. The first track, match to document, tested systems that determine if a person in a selfie is the same person pictured on an ID document. The second track, document validation, tested systems that determine if an ID is genuine or not. The third track, presentation attack detection, tested subsystems that determine if a presentation is from a legitimate or fraudulent user, like someone using a mask to impersonate someone.
Selfie Match to Document Results
The Remote Identity Validation Rally (RIVR) evaluated the current state of the industry when it comes to matching selfies to the face images on identity documents. Selfie Match to Document Systems (MTDSs) processed images from volunteers, gathered in data collection events in Maryland and California. Each volunteer’s selfies and document images were captured with three different smartphones.
MTDS performance was quantified by measuring four metrics. Failure To Extract Rates (FTXR) were separately computed for selfie and document images. False Match Rate (FMR) and False Non-Match Rate (FNMR) between volunteers’ selfie and document images was assessed at a threshold supplied with each MDTS intended to achieve a False Match Rate (FMR) of 1:10,000. Document FTXR and FNMR were calculated separately for the two available document states of issue.
The table below summarizes performance of the MTDSs. The number shown represents the “worst case” performance, the combination of smartphone and document state of issue where the MTDS had the highest error rate. Hover over each cell in the table to see the MTDS’s results disaggregated by smartphone and document state of issue. Num represents the number of errors, and Denom represents the MTDS's number of attempts. The highest error rates are bolded in the pop up.
Printable results and further information about the Selfie Match to Document track in RIVR are provided here.
| MTDS | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
Selfie FTXR |
0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% |
|
|
Document FTXR | 0% | <0.09% | 0% | 0% | <27.1% | 0% | <2.02% | 0% | <3.69% | <0.34% | 0% | <0.35% | <37.47% | <1.01% | 0% | 0% |
|
|
FNMR | <0.34% | <0.68% | <0.08% | <0.08% | <70.84% | <99.66% | <0.08% | <58.11% | <3.11% | <0.68% | <2.03% | <0.08% | <9.3% | 0% | <0.68% | <4.38% |
|
|
FMR | <0.01% | <0.02% | <0.02% | <0.01% | <0.47% | <1.66% | <0.01% | <0.01% | <0.07% | <0.01% | <0.02% | <0.01% | 0% | <0.02% | <0.01% | <0.15% |
| Legend | |||||
|---|---|---|---|---|---|
| X | Met Goal | X | Met Threshold | X | Did not meet Threshold |