Deploying modern video trackers at scale is bottlenecked by the computational cost of RGB-based object detectors. To this end, we present MVTrack, an ultrafast tracker for moving objects that operates directly on H.
Summary generated by The Flow from the publisher's feed. The full article lives at Hugging Face Trending Papers.
arXiv:2608. 10790v1 Announce Type: cross Abstract: Deploying modern video trackers at scale is bottlenecked by the computational cost of RGB-based object detectors.
arXiv:2606. 14094v1 Announce Type: cross Abstract: Conventional RGB cameras have been widely used in multi-object tracking due to their ability to capture rich appearance and semantic information.
arXiv:2607. 14898v1 Announce Type: cross Abstract: Real-time video generation demands fast decoding as much as fast denoising, yet current latent video diffusion models rely on 3D convolutional decoders that are slow and memory-intensive at high resolutions or for long video.
arXiv:2603. 24016v2 Announce Type: replace-cross Abstract: Multi-Object Tracking (MOT) has traditionally focused on a few specific categories, restricting its applicability to real-world scenarios involving diverse objects.
arXiv:2606. 23604v2 Announce Type: replace-cross Abstract: The tracking-by-detection paradigm in multi-object tracking (MOT) typically relies on static appearance descriptors to complement motion estimation.
Event cameras offer microsecond-level temporal resolution and high dynamic range for low-altitude UAV perception. However, long-range UAVs often produce sparse, fragmented, and noise-contaminated event responses, where one semantic target may appear as multiple spatially separated blobs.