Halo

Halo is a multi-camera high resolution imaging system with AI facial identification technology. Equipped with 18 megapixel machine vision cameras and specially designed LED panels, image quality was the priority when designing Halo. The images capture every detail from every angle to improve photo quality for clients. Halo matches all captured images by using AI facial identification technology. Moreover, Halo can capture high quality moving video footage that is compliant with Promat™ and Viper™ systems. This is the product of our work in offender identification and Acumé Forensic are proud to be working with two leading UK universities in facial and full body biometric research projects.

AI facial identification technology

A still image ID parade application allows a simultaneous interactive comparison of up to 12 subjects, all viewable from any angle. Halo uses multiple cameras to capture high resolution still images or video sequences. The cameras record facial images from eight different viewpoints with an additional hand held camera to record non-facial soft biometric features. This can include, tattoos and scars.

In 2007 Home Office guidance recommended that 5 custody images be taken, yet no Police Force has this capability. Todays custody images are usually low quality web cam or photo booth images. They are also usually lower quality than the prisoners photographs taken more than 30 years ago. However, the Halo imaging system has been tried and tested for millions of images. In addition, it is equipped with 18 megapixel machine vision cameras for ultimate reliability.

The Halo software interface is also designed to guide the operator through a simple record and capture process which needs minimal training. For example, a single operator is capable of recording the full series of images in under 2 minutes.

AI facial recognition

The use of Artificial Intelligence allows for identifications across many different media types, including low resolution CCTV and photographic images. The class leading AI algorithm produces 1-to-1 and 1-to-many matching of faces with no false positive identifications. Automatic enrolment learns the entire face, not just features, for more robust comparison. As a result, the comparison takes seconds with faces being compared at rates of up to 100,000 images per core.

What people are saying

"Police trial 'Big Brother' AI system that is so powerful it can identify a criminal from a brief glimpse of their ear."

Richard Gray, Daily Mail

"Having seen a marked drop in custody image quality to the point where images taken 20-30 years ago are of higher quality than those taken yesterday, combined with Home Office issued guidance in 2007 which advises that police should be capturing 5 images as a minimum, of everyone taken into custody pre charge, to date no force has been able to attain this standard. As a result, this led to the development of Halo."

Stephen Cole, Acumé Forensic

"Yesterday pioneering crime fighting technology that was developed by Acume Forensic and our academic Hassan Ugail was featured on BBC One’s Crimewatch Roadshow programme."

University of Bradford

Talk to an expert

Discuss your facial identification requirements today with one of our in-house experts.