close
close

Facial recognition back in crime-fighting toolkit for Colorado police

Facial recognition technology is returning to the Arapahoe County Sheriff's Office (ACSO), this time under a new state law that regulates the use of facial recognition in police work.

The sheriff's department had previously used FRT with biometric data from Colorado-based Rank One Computing (ROC) to identify unknown suspects by matching them to an image database. News reports from ABC Channel 7 in Denver and other sources do not specify when the previous deployment occurred — only that police use “ceased after the software provider discontinued service.”

The software provider in question was LexisNexis and the service was its product Lumen. A comprehensive draft statement of accountability for the deployment indicates that the software used an algorithm from multimodal biometrics and computer vision provider ROC Corporation, namely ROC SDK version 2.2.1.

The report addresses issues such as image quality (noting that “the core facial recognition algorithms depend primarily on the image quality of the sample and candidate images, as well as the robustness of the algorithm development process) and potential biases in algorithmic facial recognition systems.

It also sets out the policy and procedural cornerstones for the continued use of the FRT by law enforcement authorities.

“Facial recognition services offer many opportunities to increase productivity, better solve crimes, increase the effectiveness of investigations, and increase safety for citizens and members,” the policy statement states. “It is the policy of the Sheriff's Office to use facial recognition services to develop leads on unknown persons for police investigations in a manner that protects against potential misuse. This policy ensures that the use of facial recognition services by the Sheriff's Office and its members is consistent with authorized purposes while not violating anyone's privacy, civil rights, and civil liberties.”

ACSO Deputy John Bartmann says the department's initial use of facial recognition led to a number of misconceptions about what the biometric technology actually is and does. Lumen searches the Colorado Information Sharing Consortium (CISC), a government database of booking photos from law enforcement agencies across the state, to uncover possible matches that can lead to leads.

According to the accountability report, “Lumen delivers multiple results, each with a specific accuracy score generated by the ROC SDK's facial recognition algorithms. The accuracy score is intended to indicate the probability that the examination image matches a specific result.”

“There is no social media scraping, no photos from the internet or anything like that,” says Bartmann. “These are booking photos taken by the law enforcement agencies belonging to the CISC.”

The ACSO has already held two public meetings to gather feedback on facial recognition technology and is hosting a third meeting in Denver on September 12.

The impending reintroduction of facial recognition will bring its own challenges. But by now, FRT is old hat for police in Colorado. A 2020 Denver Post article points out that the Denver DMV has been fielding requests from law enforcement for FRT searches since at least 2016.

Article topics

biometric identification | biometrics | Colorado | criminal identification | facial recognition | regulation | ROC

Latest news on biometrics

During the Tencent Global Digital Ecosystem Summit, Tencent Cloud unveiled a palm verification system that uses infrared cameras to capture…

The Electoral Commission of Ghana (EC) has announced the integration of an authenticity detection feature into its biometric voter verification…

Namibia's newly proposed Civil Registration and Identification Bill has been praised for incorporating essential data protection elements. However,…

Anyone in Kuwait who fails to meet the deadlines for biometric registration must expect the suspension of government services. The Central Bank…

By Andrew Black, Managing Director, ConnectID If you work on developing digital IDs and interoperable global systems, you will be aware that…

The integration of technologies such as artificial intelligence (AI), blockchain and digital identity systems improves efficiency and inclusivity…