Zobrazeno 1 - 10
of 509
pro vyhledávání: '"Hebert, Martial"'
Despite recent significant strides achieved by diffusion-based Text-to-Image (T2I) models, current systems are still less capable of ensuring decent compositional generation aligned with text prompts, particularly for the multi-object generation. Thi
Externí odkaz:
http://arxiv.org/abs/2312.06712
Multi-task visual learning is a critical aspect of computer vision. Current research, however, predominantly concentrates on the multi-task dense prediction setting, which overlooks the intrinsic 3D world and its multi-view consistent structures, and
Externí odkaz:
http://arxiv.org/abs/2309.17450
Autor:
Keselman, Leonid, Hebert, Martial
Fast, reliable shape reconstruction is an essential ingredient in many computer vision applications. Neural Radiance Fields demonstrated that photorealistic novel view synthesis is within reach, but was gated by performance requirements for fast reco
Externí odkaz:
http://arxiv.org/abs/2308.14737
Typical black-box optimization approaches in robotics focus on learning from metric scores. However, that is not always possible, as not all developers have ground truth available. Learning appropriate robot behavior in human-centric contexts often r
Externí odkaz:
http://arxiv.org/abs/2308.04571
Large-scale datasets are essential to modern day deep learning. Advocates argue that understanding these methods requires dataset transparency (e.g. "dataset curation, motivation, composition, collection process, etc..."). However, almost no one has
Externí odkaz:
http://arxiv.org/abs/2306.14035
Publikováno v:
CVPR 2023
Object discovery -- separating objects from the background without manual labels -- is a fundamental open challenge in computer vision. Previous methods struggle to go beyond clustering of low-level cues, whether handcrafted (e.g., color, texture) or
Externí odkaz:
http://arxiv.org/abs/2303.15555
Autor:
Keselman, Leonid, Hebert, Martial
Many practitioners in robotics regularly depend on classic, hand-designed algorithms. Often the performance of these algorithms is tuned across a dataset of annotated examples which represent typical deployment conditions. Automatic tuning of these s
Externí odkaz:
http://arxiv.org/abs/2303.07434
Orientation estimation is the core to a variety of vision and robotics tasks such as camera and object pose estimation. Deep learning has offered a way to develop image-based orientation estimators; however, such estimators often require training on
Externí odkaz:
http://arxiv.org/abs/2211.11182
Near-Periodic Patterns (NPP) are ubiquitous in man-made scenes and are composed of tiled motifs with appearance differences caused by lighting, defects, or design elements. A good NPP representation is useful for many applications including image com
Externí odkaz:
http://arxiv.org/abs/2208.12278
Autor:
Keselman, Leonid, Hebert, Martial
Differentiable renderers provide a direct mathematical link between an object's 3D representation and images of that object. In this work, we develop an approximate differentiable renderer for a compact, interpretable representation, which we call Fu
Externí odkaz:
http://arxiv.org/abs/2207.10606