One of today’s biggest challenges in fighting cybercrime is the undetected production and spread of Child Sexual Abuse (CSA) content.

While Hash Databases or PhotoDNA are widely used to find previously known and registered images, the need for tools robustly detecting inappropriate media data lies in the simple ability to trick these methods by just modifying images, which makes them much more difficult to detect through database solutions.

For T3K’s Founder and CEO Felix Klier, the fight against CSA is one of the biggest goals of his company: “Anything we can do to help find victims faster and to protect children from predators, is a step in the right direction. That’s why the development of relevant Object Recognition has been a big focus of T3K in the past years and I’m very proud to introduce it to the market now, offering our revolutionary approach to anybody fighting CSA.”

In recent years, due to the issues mentioned above, the development of AI-based tools to find new and unknown CSA images has become a major focus. However, the majority of these tools are not state-of-the-art in respect to current research, e.g. working with image recognition that is trained to scan the full picture for contents, which often results in biased data or detection of image backgrounds, ultimately leading to high false-positive rates.

T3K’s Object Recognition technology IVi (Image and Video Identifier Engine) takes a novel approach to address this issue by incorporating a two-stage scanning process for CSA – the first stage defining that the image or video shows pornography and the second stage identifying children or youth in the media. IVi is already successfully used in T3K LEAP (Law Enforcement Analytical Platform).

“We are extremely happy to bring LEAP and CORE to the daunting challenge of detecting Child Sexual Abuse images and videos. This new detection technology and performance benchmark should greatly help those officers and security analysts burdened by the manual review of this emotionally stressful material. This level and speed of detection is what I would have liked to have used when I was a cybercrimes investigator.”, says Jim Plitt, CEO of T3K Inc. (USA). “But adding generic AI to a digital forensics tool just isn’t enough for this critical law enforcement category. You have to have the right high-performance AI engine to make good on that important promise. IVi is that engine, an amazing piece of technology far above the common label of its genre – AI.”

The IVi technology is part of our products LEAP and CORE.

As one part of the LEAP platform, IVi is used to find inappropriate images and videos of children and youth. With the process as described above, LEAP manages to achieve unseen performance and detection rates, increasing the true-positive rate dramatically, while at the same time significantly reducing false-positives, through the two-stage approach and elimination of Background Factoring.

All this is especially true for Video Analysis, where IVi sets a new detection benchmark.

We are very proud to launch our new product, CORE (Content Review Engine), which is an image and video classifying tool that is also operated by IVi. It stands independently from the LEAP platform and can be integrated into existing content screening workflows at corporations, data hosting providers or government agencies, to detect CSA content on internal storage or hosted data.

Similar to LEAP, CORE detects more inappropriate data with less false-positives through the IVi technology.

For more information on LEAP, CORE or T3K.AI, please contact us at office@t3k.ai