This paper develops a unified framework for image-to-image translation based on
conditional diffusion models and evaluates this framework on four challenging image-to …

Numerous task-specific variants of conditional generative adversarial networks have been
developed for image completion. Yet, a serious limitation remains that all existing algorithms …

Advancing image inpainting is challenging as it requires filling user-specified regions for
various intents, such as background filling and object synthesis. Existing approaches focus …

Abstract Vision-and-Language Navigation requires the agent to follow language instructions
to navigate through 3D environments. One main challenge in Vision-and-Language …

We present a method for learning to generate unbounded flythrough videos of natural
scenes starting from a single view. This capability is learned from a collection of single …

We introduce the problem of perpetual view generation-long-range generation of novel
views corresponding to an arbitrarily long camera trajectory given a single image. This is a …

Recent advancements in differentiable rendering and 3D reasoning have driven exciting
results in novel view synthesis from a single image. Despite realistic results, methods are …

Despite increasingly realistic image quality, recent 3D image generative models often
operate on 3D volumes of fixed extent with limited camera motions. We investigate the task …

Diffusion models (DMs) have recently gained attention with state-of-the-art performance in
text-to-image synthesis. Abiding by the tradition in deep learning, DMs are trained and …

Generative models operate at fixed resolution, even though natural images come in a variety
of sizes. As high-resolution details are downsampled away and low-resolution images are …