CORE (Content Review Engine) is an image and video classifier that screens media data for illegal and inappropriate content. As a pure API solution allowing easy integration into existing infrastructure, CORE supports digital forensics, eDiscovery and other review platforms with enhanced media analytics. CORE can be deployed on premise, in the cloud or as part of a hybrid setup.

Advantages:

API-level access to T3K’s machine learning technology provides the ability to detect contents in data at scale, identifying bad actors in online networks, through hosted data, or through seized evidence. Users can perform quick and complete scans of the data CORE is directed at.

●  Running on internal or external online networks

●  Scanning millions of media contents per day

●  Detecting new and self-produced content not found in hash databases

●  Extremely low false positive ratio

●  Integrates into third party tools or directly into networks, and scales easily according to data volume

●  Used globally by leading law enforcement authorities for detection of CSAM (child sexual abuse material, TVEC (terrorist and violent extremist content, financial crime, and many more.

Features:

CORE’s screening capabilities accelerate investigations which involve large amounts of seized digital devices. The models are constantly evolving and are currently generating matches on the following recognition classes:

●  Child sexual abuse material (CSAM), incl. facial age and gender estimation

●  Identity cards: passports and cards potentially displaying an identity

●  Motor Vehicles

●  Military equipment: guns, military uniforms, armoured vehicles

●  Screenshots, documents and maps, handwriting

●  Terrorist and violent extremist content (TVEC): masked individuals, ISIS flags and different logos of propaganda media outlets

● CORE can conduct multi-lingual OCR (Latin, Arabic and Cyrillic alphabets), to detect search terms from uploaded watchlists.

●  It offers hash and PhotoDNA matching and checking against the ProjectVIC CSAM database.

●  CORE provides face recognition from uploaded reference images.

●  CORE detects key elements of visual 2D patterns to recognize logos, art, bank notes, tattoos, etc. from uploaded reference images.

CORE Setup:

Seamless integration into existing backend infrastructure or existing platforms like Griffeye DI Pro/CS Ops/CS Enterprise, Magnet AUTOMATE/AXIOM. Media objects are directed to the CORE server during or after processing in the existing platform, automatically classified based on the chosen Al classifiers, and tagged for review in the integrating platform’s frontend. Media data is also “indexed”, allowing real-time search using natural language keywords.