Contributor Revolutionizing Computer Vision: OpenFilter with Andrew Smith
Jun 18, 2025
Andrew Smith, the CTO of Plainsight and expert in computer vision, discusses the revolutionary OpenFilter framework. He highlights the challenges of deploying computer vision in edge environments and how experiences from Google shaped his vision. Andrew shares the three key issues he aimed to resolve by open-sourcing the project and presents compelling use cases for OpenFilter. The conversation emphasizes the importance of standardization and interoperability in enhancing data management and real-time processing in various industries.
AI Snips
Chapters
Transcript
Episode notes
OpenFilter Standardizes Computer Vision
- OpenFilter enables building modular computer vision pipelines that run uniformly on cloud, edge, and local devices.
- This standardizes development and deployment, creating predictability in complex computer vision workflows.
Tackling Diverse Deployment Challenges
- Computer vision use cases vary by scale: automotive and retail can invest heavily while smaller or remote sites struggle.
- OpenFilter addresses the challenge of diverse environments by enabling consistent algorithms across different hardware and connectivity constraints.
Filtering Video Data at Multiple Stages
- Data gravity and cost prohibit sending all video data to the cloud or running GPUs at each edge camera.
- OpenFilter introduces filtering stages at camera, edge, and cloud to reduce data transmission and computational expense.