CMU Panoptic
PulseAugur coverage of CMU Panoptic — every cluster mentioning CMU Panoptic across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
New method enables real-time 4D reconstruction of fast-moving objects
Researchers have introduced FastFlowGS, a novel method for streaming 4D reconstruction of fast-moving subjects using a limited set of external cameras. This technique is benchmarked on Monaco4D, a new photorealistic dat…
-
New MEOM method enhances 3D human pose estimation accuracy
Researchers have developed a new method called Multi-View Expected-OKS Maximization (MEOM) for improving 3D human pose estimation from multi-view 2D keypoints. MEOM addresses limitations in conventional methods by utili…
-
New VQ-VAD framework uses discrete motion for video anomaly detection
Researchers have developed VQ-VAD, a new framework for human-centric video anomaly detection that learns discrete motion representations. This approach adapts Vector-Quantized GAN (VQ-GAN) to create a codebook of normal…
-
New AI models InterMesh and Anny-Fit advance 3D human pose and shape recovery
Researchers have developed InterMesh, a new framework for multi-person human mesh recovery that explicitly incorporates human-environment interaction information. This approach enhances pose and shape estimation by enri…