Case Study: FLIR achieves synthetic data training for thermal AI with CVEDIA

FLIR achieves 100% synthetic data training with CVEDIA

FLIR, a world-renowned designer and manufacturer of thermal cameras, faced a major obstacle in gathering sufficient real-world training data for their computer vision algorithms. This high data requirement made developing many of their use cases impossible. FLIR's CTO met with CVEDIA and decided to invest in the company to leverage its synthetic data technology.

CVEDIA collaborated with FLIR to design a thermal synthetic data pipeline that could train neural networks without the collection of more real data. This solution was integrated into FLIR's engineering pipeline, solving the problem of domain adaptation in 2019. As a result, FLIR deployed the technology across multiple applications, enhancing their AI roadmap and maintaining their position as a market leader. This included launching a deer collision detection algorithm for ADAS and supporting defense applications, all powered by CVEDIA's synthetic data.

FLIR