Computer vision techniques, in conjunction with acquisition through remote cameras and
unmanned aerial vehicles (UAVs), offer promising non-contact solutions to civil infrastructure …

Synthetic datasets, for which we propose the term synthsets, are not a novelty but have
become a necessity. Although they have been used in computer vision since 1989, hel** …

We introduce PointOdyssey, a large-scale synthetic dataset, and data generation framework,
for the training and evaluation of long-term fine-grained tracking algorithms. Our goal is to …

Conditional image-to-video (cI2V) generation aims to synthesize a new plausible video
starting from an image (eg, a person's face) and a condition (eg, an action class label like …

Effectively extracting inter-frame motion and appearance information is important for video
frame interpolation (VFI). Previous works either extract both types of information in a mixed …

Data is the driving force of machine learning, with the amount and quality of training data
often being more important for the performance of a system than architecture and training …

Prevailing video frame interpolation algorithms, that generate the intermediate frames from
consecutive inputs, typically rely on complex model architectures with heavy parameters or …

We present VIDIM a generative model for video interpolation which creates short videos
given a start and end frame. In order to achieve high fidelity and generate motions unseen in …

Efficient reasoning about the semantic, spatial, and temporal structure of a scene is a crucial
prerequisite for autonomous driving. We present NEural ATtention fields (NEAT), a novel …

We present a frame interpolation algorithm that synthesizes an engaging slow-motion video
from near-duplicate photos which often exhibit large scene motion. Near-duplicates …