Probes are small networks that predict properties of underlying data fro...
Denoising diffusion probabilistic models have transformed image generati...
Humans excel at acquiring knowledge through observation. For example, we...
Inverse rendering of an object under entirely unknown capture conditions...
How important are training details and datasets to recent optical flow m...
We present a frame interpolation algorithm that synthesizes multiple
int...
Image super-resolution (SR) is a fast-moving field with novel architectu...
Aggressive data augmentation is a key component of the strong generaliza...
Remarkable progress has been made in 3D reconstruction of rigid structur...
Synthetic datasets play a critical role in pre-training CNN models for
o...
Prototype learning is extensively used for few-shot segmentation. Typica...
In this paper, we address the problem of building dense correspondences
...
End-to-end deep learning methods have advanced stereo vision in recent y...
One central question for video action recognition is how to model motion...
We introduce a compact network for holistic scene flow estimation, calle...
We present a simple and effective image super-resolution algorithm that
...
Despite the long history of image and video stitching research, existing...
Learning to synthesize high frame rate videos via interpolation requires...
Convolutions are the fundamental building block of CNNs. The fact that t...
To date, top-performing optical flow estimation methods only take pairs ...
We investigate two crucial and closely related aspects of CNNs for optic...
Superpixels provide an efficient low/mid-level representation of image d...
We address the unsupervised learning of several interconnected problems ...
In this paper, we propose a general dual convolutional neural network
(D...
Estimation of 3D motion in a dynamic scene from a temporal pair of image...
We present a network architecture for processing point clouds that direc...
We study domain-specific video streaming. Specifically, we target a stre...
Given two consecutive frames, video interpolation aims at generating
int...
We present a compact but effective CNN model for optical flow, called
PW...
Given two consecutive frames from a pair of stereo cameras, 3D scene flo...
Existing optical flow methods make generic, spatially homogeneous,
assum...