Thank you for your excellent work on SFSORT! 🙌
I'm currently deploying on NVIDIA Jetson edge devices, and I found that DeepStream (based on GStreamer) enables direct GPU hardware encoding/decoding, which significantly improves inference efficiency and reduces latency.
Question
I'd like to ask if there are any plans to add DeepStream plugin support for SFSORT?
Reference Implementation
For your reference, ByteTrack has already implemented a DeepStream integration:
🔗 https://github.com/FoundationVision/ByteTrack/tree/main/deploy/DeepStream
Potential Benefits
- ✅ Real-time tracking with low latency on Jetson series (Nano / Xavier / Orin)
- ✅ Zero-copy GPU pipeline: decoder → inference → tracker → encoder
- ✅ Better resource efficiency for edge deployment scenarios
I'd be happy to help with testing or contribution if needed. Thanks for considering this feature!
Thank you for your excellent work on SFSORT! 🙌
I'm currently deploying on NVIDIA Jetson edge devices, and I found that DeepStream (based on GStreamer) enables direct GPU hardware encoding/decoding, which significantly improves inference efficiency and reduces latency.
Question
I'd like to ask if there are any plans to add DeepStream plugin support for SFSORT?
Reference Implementation
For your reference, ByteTrack has already implemented a DeepStream integration:
🔗 https://github.com/FoundationVision/ByteTrack/tree/main/deploy/DeepStream
Potential Benefits
I'd be happy to help with testing or contribution if needed. Thanks for considering this feature!