Case Study: a leading technology company enhances threat detection algorithms with Twine's custom video datasets

Enhancing Threat Detection Algorithms through Custom Video Datasets

Twine collaborated with a leading technology company that needed to enhance its threat detection algorithms. The challenge was to create a custom video dataset for a face detection project, requiring videos of actors enacting violent scenes that balanced realism with algorithmic performance.

The solution from Twine involved producing 25 custom videos per group, shot with static cameras from an elevated perspective to mimic authentic security footage. The client reported being highly satisfied with the quality and effectiveness of the initial batch, which allowed them to immediately begin the annotation process and make significant progress in improving their algorithms. Twine exceeded expectations, leading to plans for further collaboration.