Page 53 - WEB 2017 Annual Report_Float

CASE STUDY: FACIAL RECOGNITION IDENTIFIES THEFT SUSPECT

STORE SECURITY FOOTAGE WCSO BOOKING PHOTO SOCIAL MEDIA PHOTO

• Suspect steals almost $1,000 from a local big box by Adzima and connects a public profile picture to a
hardware store using self-checkout to scan items, but suspect
leaves without paying • Rekognition and detectives identify facial similarities
• Photo of suspect is captured by store security cameras across all three photos - eyes, mouth, similar chin and
and shared with WCSO ears
• Rekognition searches through 300,000+ booking • Suspect also wearing the same sweatshirt in social
photos and returned a match within seconds media photo as the theft photo
FACIAL RECOGNITION TECHNOLOGY • Adzima provides a detective with the suspect’s name • Following possible positive identification, detectives
are able to proceed with investigation and locate
for further investigation
ENHANCES INVESTIGATIONS • Detective searches social media and identifies a suspect for questioning
Facebook page associated with the name provided
Often in an investigation a security camera or Adzima began to test out Amazon Rekognition
cellphone photo may catch a small glimpse of a technology as a way to quickly access and scan
suspect or someone connected to a case. What booking photos. The technology works similarly
may be a blurred or distorted image is regularly to programs personal photographers may use
shared with local law enforcement agencies to to organize and catalog family photos. Images,
determine if anyone can identify the individual. in this case booking photos, are analyzed and LAW ENFORCEMENT TECHNOLOGY
Until 2017, this was (and in some cases still the service identifies individual faces that
is) a consistent practice. While this system is are a possible match. Adzima uploaded and Law Enforcement Technology (LET) is an Help Me Home is a voluntary
sometimes effective, it is far from efficient. indexed the WCSO database of over 300,000 information technology unit working directly enrollment program for adults and
booking photos, translating to roughly 200,000 for the Sheriff. LET consists of a small team children who may become lost and
Chris Adzima, Senior Information System Analyst individuals. As part of the project policies were of information systems analysts who directly have difficulty communicating vital
with WCSO’s Law Enforcement Technology developed to define when and how the system support law enforcement related technology information about themselves.
(LET) group knew there had to be a better way. would be used in criminal cases and how it will Those who are registered in this
Over 17,000 people are booked into the WCSO be eventually used to identify at-risk individuals matters; performing data analysis, supporting program have their identification,
jail every year. Each booking requires a photo registered in the WCSO Help Me Home database. legacy systems, developing new applications, photo and pertinent information
to document the individual’s appearance. In and researching new technology that adds placed in a database managed by
the Sheriff’s Office. This database
addition to front and side profile pictures, the Ready access to photos and the ability to quickly value to the Sheriff’s Office operations. can be accessed by any law
jail also photographs significant scars, marks and identify possible matches provides a boost to enforcement officer in Washington
tattoos. The photographs are accessible to law the investigative process – even expediting County via laptops in their patrol
enforcement but require a specific search process an investigation by providing supplemental vehicles.
to access. information to a deputy or detective.

49 | WCSO Patrol Division WCSO 2017 Annual Report | 50