Wonmin Byeon
Researcher at NVIDIA
In recent years, video generation has become a prominent generative tool...
We propose a novel framework and a solution to tackle the continual lear...
We present a novel framework, Localized Image Stylization with Audio (LI...
Recent successes suggest that an image can be manipulated by a text prom...
The recent success in StyleGAN demonstrates that pre-trained StyleGAN la...
Physics-informed neural networks allow models to be trained by physical ...
The recent success of the generative model shows that leveraging the
mul...
Enforcing orthogonality in neural networks is an antidote for gradient
v...
A major challenge for physically unconstrained gaze estimation is acquir...
We present SimNet, an AI-driven multi-physics simulation framework, to
a...
Although deep learning-based methods have dominated stereo matching
lead...
Higher-order Recurrent Neural Networks (RNNs) are effective for long-ter...
In many cases, especially with medical images, it is prohibitively
chall...
We introduce a data-driven forecasting method for high dimensional, chao...
This paper proposes a new neural network design for unsupervised learnin...
Convolutional Neural Networks (CNNs) can be shifted across 2D images or ...