Zobrazeno 21 - 30
of 258
pro vyhledávání: '"Lucey, Simon"'
In this paper we present a new method for deformable NeRF that can directly use optical flow as supervision. We overcome the major challenge with respect to the computationally inefficiency of enforcing the flow constraints to the backward deformatio
Externí odkaz:
http://arxiv.org/abs/2303.16333
Modelling dynamical systems is an integral component for understanding the natural world. To this end, neural networks are becoming an increasingly popular candidate owing to their ability to learn complex functions from large amounts of data. Despit
Externí odkaz:
http://arxiv.org/abs/2303.05728
We introduce a general theoretical framework, designed for the study of gradient optimisation of deep neural networks, that encompasses ubiquitous architecture choices including batch normalisation, weight normalisation and skip connections. Our fram
Externí odkaz:
http://arxiv.org/abs/2210.05371
Autor:
Dabhi, Mosam, Wang, Chaoyang, Clifford, Tim, Jeni, Laszlo Attila, Fasel, Ian R., Lucey, Simon
Labeling articulated objects in unconstrained settings have a wide variety of applications including entertainment, neuroscience, psychology, ethology, and many fields of medicine. Large offline labeled datasets do not exist for all but the most comm
Externí odkaz:
http://arxiv.org/abs/2210.01721
The role of quantization within implicit/coordinate neural networks is still not fully understood. We note that using a canonical fixed quantization scheme during training produces poor performance at low-rates due to the network weight distributions
Externí odkaz:
http://arxiv.org/abs/2209.01019
Autor:
Ramasinghe, Sameera, MacDonald, Lachlan, Farazi, Moshiur, Saratchandran, Hemanth, Lucey, Simon
Characterizing the remarkable generalization properties of over-parameterized neural networks remains an open problem. In this paper, we promote a shift of focus towards initialization rather than neural architecture or (stochastic) gradient descent
Externí odkaz:
http://arxiv.org/abs/2206.08558
It is well noted that coordinate-based MLPs benefit -- in terms of preserving high-frequency information -- through the encoding of coordinate positions as an array of Fourier features. Hitherto, the rationale for the effectiveness of these positiona
Externí odkaz:
http://arxiv.org/abs/2205.08987
Despite Neural Radiance Fields (NeRF) showing compelling results in photorealistic novel views synthesis of real-world scenes, most existing approaches require accurate prior camera poses. Although approaches for jointly recovering the radiance field
Externí odkaz:
http://arxiv.org/abs/2204.05735
We address the problem of map sparsification for long-term visual localization. For map sparsification, a commonly employed assumption is that the pre-build map and the later captured localization query are consistent. However, this assumption can be
Externí odkaz:
http://arxiv.org/abs/2203.15182
We show that typical implicit regularization assumptions for deep neural networks (for regression) do not hold for coordinate-MLPs, a family of MLPs that are now ubiquitous in computer vision for representing high-frequency signals. Lack of such impl
Externí odkaz:
http://arxiv.org/abs/2202.00790