Abstract
Footage from security cameras deployed within an organization often includes uniformed employees of the organization, who generally present no security threats. Surfacing the presence of uniformed personnel in security footage can amount to false alarms. This disclosure describes techniques that can identify persons wearing authorized uniforms within security camera footage. Per the techniques, generative artificial intelligence (gen-AI) is used to synthesize images of people wearing various uniforms authorized by the organization, in various poses. Embeddings of the gen-AI images are created and compared with those corresponding to real persons detected in the security camera footage to determine if a detected person is likely a uniformed employee. The techniques can be used to add context to person detection by identifying a detected person to likely be a uniformed employee (or not). By thus adding context, false alarms are averted. Search of the security footage can easily be narrowed to persons or events of genuine security interest.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
NA, "Face Agnostic Recognition of Uniformed Personnel Using Generative Artificial Intelligence", Technical Disclosure Commons, ()
https://www.tdcommons.org/dpubs_series/9513