Fenon Robotics released a video quality control tool fgate

Company fgate presented a solution for automated quality control of first-person video streams. The tool is oriented towards real-time operation and provides reproducible content checking with the generation of typed reports.

The system uses hardware acceleration via TensorRT and NVDEC, which allows for processing video streams directly on the GPU. To ensure data integrity, each report is accompanied by hash verification, which eliminates the possibility of substituting or distorting the verification results. The solution's architecture is built on a local Docker backend, which provides process isolation and stable operation in various environments.

The project is aimed at automating QC processes (Quality Control) in tasks that require strict verification of video data. Developers confirmed that a cloud version of the service will appear in the upcoming updates, expanding scalability capabilities for large projects. The source code is available in the project's repository on GitHub.

The implementation of such tools allows for standardizing video content checking in computer vision and robotics tasks. The use of NVIDIA hardware libraries minimizes delays during analysis, which is critical for systems working with streaming data 24/7.