CVMay 10, 2025

Video Dataset Condensation with Diffusion Models

arXiv:2505.06670v11 citationsh-index: 9
Originality Incremental advance
AI Analysis

This addresses the computational resource challenge for researchers and practitioners working with large video datasets, representing an incremental advance in dataset distillation specifically for videos.

The paper tackles the problem of reducing computational demands for video deep learning by developing a video dataset distillation method that generates compact synthetic datasets, achieving performance improvements of up to 10.61% over state-of-the-art methods on benchmark datasets.

In recent years, the rapid expansion of dataset sizes and the increasing complexity of deep learning models have significantly escalated the demand for computational resources, both for data storage and model training. Dataset distillation has emerged as a promising solution to address this challenge by generating a compact synthetic dataset that retains the essential information from a large real dataset. However, existing methods often suffer from limited performance and poor data quality, particularly in the video domain. In this paper, we focus on video dataset distillation by employing a video diffusion model to generate high-quality synthetic videos. To enhance representativeness, we introduce Video Spatio-Temporal U-Net (VST-UNet), a model designed to select a diverse and informative subset of videos that effectively captures the characteristics of the original dataset. To further optimize computational efficiency, we explore a training-free clustering algorithm, Temporal-Aware Cluster-based Distillation (TAC-DT), to select representative videos without requiring additional training overhead. We validate the effectiveness of our approach through extensive experiments on four benchmark datasets, demonstrating performance improvements of up to \(10.61\%\) over the state-of-the-art. Our method consistently outperforms existing approaches across all datasets, establishing a new benchmark for video dataset distillation.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes