Zobrazeno 1 - 10
of 30
pro vyhledávání: '"Sekikawa, Yusuke"'
When a camera travels across a 3D world, only a fraction of pixel value changes; an event-based camera observes the change as sparse events. How can we utilize sparse events for efficient recovery of the camera pose? We show that we can recover the c
Externí odkaz:
http://arxiv.org/abs/2304.04559
In this paper, we present an end-to-end unsupervised anomaly detection framework for 3D point clouds. To the best of our knowledge, this is the first work to tackle the anomaly detection task on a general object represented by a 3D point cloud. We pr
Externí odkaz:
http://arxiv.org/abs/2304.03420
We propose an action-conditional human motion generation method using variational implicit neural representations (INR). The variational formalism enables action-conditional distributions of INRs, from which one can easily sample representations to g
Externí odkaz:
http://arxiv.org/abs/2203.13694
We present a novel framework of motion tracking from event data using implicit expression. Our framework use pre-trained event generation MLP named implicit event generator (IEG) and does motion tracking by updating its state (position and velocity)
Externí odkaz:
http://arxiv.org/abs/2111.03824
Autor:
Sekikawa, Yusuke, Suzuki, Teppei
Aiming at drastic speedup for point-feature embeddings at test time, we propose a new framework that uses a pair of multi-layer perceptrons (MLP) and a lookup table (LUT) to transform point-coordinate inputs into high-dimensional features. When compa
Externí odkaz:
http://arxiv.org/abs/2011.09852
PointNet, which is the widely used point-wise embedding method and known as a universal approximator for continuous set functions, can process one million points per second. Nevertheless, real-time inference for the recent development of high-perform
Externí odkaz:
http://arxiv.org/abs/2007.15855
Autor:
Sekikawa, Yusuke, Suzuki, Teppei
Aiming at a drastic speedup for point-data embeddings at test time, we propose a new framework that uses a pair of multi-layer perceptron (MLP) and look-up table (LUT) to transform point-coordinate inputs into high-dimensional features. When compared
Externí odkaz:
http://arxiv.org/abs/1912.00790
Event cameras are bio-inspired vision sensors that mimic retinas to asynchronously report per-pixel intensity changes rather than outputting an actual intensity image at regular intervals. This new paradigm of image sensor offers significant potentia
Externí odkaz:
http://arxiv.org/abs/1812.07045
Akademický článek
Tento výsledek nelze pro nepřihlášené uživatele zobrazit.
K zobrazení výsledku je třeba se přihlásit.
K zobrazení výsledku je třeba se přihlásit.
Autor:
Nagata, Jun1 (AUTHOR), Sekikawa, Yusuke2 (AUTHOR), Hara, Kosuke2 (AUTHOR), Aoki, Yoshimitsu1 (AUTHOR) aoki@elec.keio.ac.jp
Publikováno v:
Electronics & Communications in Japan. Apr2020, Vol. 103 Issue 1-4, p19-25. 7p.