


LAW ENFORCEMENT ANALYTICS PLATFORM
LEAP is an automated tool for quick analysis of data from mobile devices or other sources. It provides automated media classification to detect illicit content. It can analyse communication patterns, verify identities and countries of origin, and quickly screen persons of interest.
01
Input-agnostic
LEAP has been designed to ingest the common mobile extraction formats: Cellebrite UFDR, MSAB Extended XML, Oxygen Forensics XML, Grayshift Graykey ZIP, MobilEdit XML and UFDR, .E01 forensic images and Raw Disk Images
03
Easy to use
LEAP is designed to be simple to use and leverages a web-based GUI allowing access on multiple device and platform types without the need to install a client on each one
05
Rapid access to key evidence
LEAP’s powerful media analysis engine is capable of rapidly screening thousands of images, videos and documents and uses machine learning technology to identify relevant content
02
Automated reporting
Automatic reporting and aggregated data allows investigators to make decisions faster and progress their investigation without waiting for manual analysis and reporting processes
04
Best-in-class AI Classifiers
LEAP employs powerful Deep Learning architecture utilizing object recognition developed and trained by our team of experts. Automated analysis of picture, video and document files allows the detection of over 20 specific object classes
06
Cloud ready
LEAP is fully containerised, allowing for easy adoption into common cloud and edge environments. Due to the modular containerized setup, LEAP in the cloud can be scaled horizontally and vertically.
System Requirements
General Requirements:
-
GPU: The use of powerful GPUs is highly recommended when using image-object recognition features
-
OS: Ubuntu Linux 20.04 LTS (Ubuntu Linux 22.04 LTS will be supported soon)
-
Storage: LEAP & OS itself need approximately 400GB (including backups); the storage solution's size should be aligned to the amount of data to be processed and their storage times. LEAP’s performance benefits immensely from using NVMe SSDs for storing extractions
The following requirements are meant as reference point for potential system environments:
Small solution
These are the minimum requirements for running between 5 to 10 mid-size extractions a day using image-object recognition:
-
CPU: At least 10th Generation Intel Core i7, AMD Ryzen 7 3000-series or comparable
-
RAM: 128GB
-
GPU: NVIDIA Quadro RTX 4000 or better CUDA-enabled card
Mid to large size solution
For high throughput and time-critical environments, T3K recommends the following minimum requirements:
-
CPU: 2x Intel Xeon Gold 6142/1x AMD Ryzen Threadripper 2990WX or better CPUs with at least 32 cores/64 threads (2.8GHz base/3.7GHz turbo per core)
-
RAM: 192GB DDR4
-
GPU: NVIDIA RTX A6000 or better CUDA-enabled card