LAVIB: A Large-scale Video Interpolation Benchmark

TL;DR

LAVIB is a comprehensive large-scale benchmark dataset for evaluating video frame interpolation methods, featuring diverse high-resolution videos and multiple challenging splits to advance research in low-level video tasks.

Contribution

The paper introduces LAVIB, a new extensive dataset for video frame interpolation, with diverse metrics and challenging splits, addressing gaps in current datasets.

Findings

Includes 283K clips from 17K videos, totaling 77.6 hours.

Provides diverse metrics for video analysis.

Creates challenging OOD splits for robust evaluation.

Abstract

This paper introduces a LArge-scale Video Interpolation Benchmark (LAVIB) for the low-level video task of Video Frame Interpolation (VFI). LAVIB comprises a large collection of high-resolution videos sourced from the web through an automated pipeline with minimal requirements for human verification. Metrics are computed for each video's motion magnitudes, luminance conditions, frame sharpness, and contrast. The collection of videos and the creation of quantitative challenges based on these metrics are under-explored by current low-level video task datasets. In total, LAVIB includes 283K clips from 17K ultra-HD videos, covering 77.6 hours. Benchmark train, val, and test sets maintain similar video metric distributions. Further splits are also created for out-of-distribution (OOD) challenges, with train and test splits including videos of dissimilar attributes.