A single-camera pipeline combining YOLOv11 detection with a real-time TensorRT-optimized DSTT inpainting model demonstrates object-level privacy redaction in MR collaboration at over 20 fps, with the caveat that depth ghosting remains.
Richter, Vibhav Vineet, Stefan Roth, and Vladlen Koltun
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A Real-Time Diminished Reality Approach to Privacy in MR Collaboration
A single-camera pipeline combining YOLOv11 detection with a real-time TensorRT-optimized DSTT inpainting model demonstrates object-level privacy redaction in MR collaboration at over 20 fps, with the caveat that depth ghosting remains.