Zobrazeno 1 - 10
of 62
pro vyhledávání: '"Honari, Sina"'
Occlusions remain one of the key challenges in 3D body pose estimation from single-camera video sequences. Temporal consistency has been extensively used to mitigate their impact but the existing algorithms in the literature do not explicitly model t
Externí odkaz:
http://arxiv.org/abs/2402.11036
Vision transformers have emerged as powerful tools for many computer vision tasks. It has been shown that their features and class tokens can be used for salient object segmentation. However, the properties of segmentation transformers remain largely
Externí odkaz:
http://arxiv.org/abs/2212.14397
Autor:
Beckham, Christopher, Weiss, Martin, Golemo, Florian, Honari, Sina, Nowrouzezahrai, Derek, Pal, Christopher
Different types of mental rotation tests have been used extensively in psychology to understand human visual reasoning and perception. Understanding what an object or visual scene would look like from another viewpoint is a challenging problem that i
Externí odkaz:
http://arxiv.org/abs/2212.01639
Analyzing and training 3D body posture models depend heavily on the availability of joint labels that are commonly acquired through laborious manual annotation of body joints or via marker-based joint localization using carefully curated markers and
Externí odkaz:
http://arxiv.org/abs/2211.12829
Publikováno v:
IEEE Robotics and Automation Letters ( Volume: 8, Issue: 4, April 2023, Pages: 2150-2157)
While road obstacle detection techniques have become increasingly effective, they typically ignore the fact that, in practice, the apparent size of the obstacles decreases as their distance to the vehicle increases. In this paper, we account for this
Externí odkaz:
http://arxiv.org/abs/2210.01779
Supervised approaches to 3D pose estimation from single images are remarkably effective when labeled data is abundant. However, as the acquisition of ground-truth 3D labels is labor intensive and time consuming, recent attention has shifted towards s
Externí odkaz:
http://arxiv.org/abs/2203.15865
Autor:
Davydov, Andrey, Remizova, Anastasia, Constantin, Victor, Honari, Sina, Salzmann, Mathieu, Fua, Pascal
The Skinned Multi-Person Linear (SMPL) model can represent a human body by mapping pose and shape parameters to body meshes. This has been shown to facilitate inferring 3D human pose and shape from images via different learning models. However, not a
Externí odkaz:
http://arxiv.org/abs/2112.04203
Relating animal behaviors to brain activity is a fundamental goal in neuroscience, with practical applications in building robust brain-machine interfaces. However, the domain gap between individuals is a major issue that prevents the training of gen
Externí odkaz:
http://arxiv.org/abs/2112.01176
A fundamental goal in neuroscience is to understand the relationship between neural activity and behavior. For example, the ability to extract behavioral intentions from neural data, or neural decoding, is critical for developing effective brain mach
Externí odkaz:
http://arxiv.org/abs/2111.14595
Autor:
Chan, Robin, Lis, Krzysztof, Uhlemeyer, Svenja, Blum, Hermann, Honari, Sina, Siegwart, Roland, Fua, Pascal, Salzmann, Mathieu, Rottmann, Matthias
State-of-the-art semantic or instance segmentation deep neural networks (DNNs) are usually trained on a closed set of semantic classes. As such, they are ill-equipped to handle previously-unseen objects. However, detecting and localizing such objects
Externí odkaz:
http://arxiv.org/abs/2104.14812