Learning Compact Video Representations for Efficient Long-form Video Understanding in Large Multimodal Models

TL;DR

This paper presents a novel end-to-end framework combining adaptive sampling and compression techniques to improve long-form video understanding with large multimodal models, effectively handling large volumes of redundant video data.

Contribution

It introduces an adaptive video sampler and a spatiotemporal compressor integrated with multimodal LLMs, enabling efficient processing of lengthy videos with high information retention.

Findings

Achieves high compression rates while preserving key information.

Demonstrates superior performance on long-form video benchmarks.

Effectively manages large, redundant video sequences.

Abstract

With recent advancements in video backbone architectures, combined with the remarkable achievements of large language models (LLMs), the analysis of long-form videos spanning tens of minutes has become both feasible and increasingly prevalent. However, the inherently redundant nature of video sequences poses significant challenges for contemporary state-of-the-art models. These challenges stem from two primary aspects: 1) efficiently incorporating a larger number of frames within memory constraints, and 2) extracting discriminative information from the vast volume of input data. In this paper, we introduce a novel end-to-end schema for long-form video understanding, which includes an information-density-based adaptive video sampler (AVS) and an autoencoder-based spatiotemporal video compressor (SVC) integrated with a multimodal large language model (MLLM). Our proposed system offers two major advantages: it adaptively and effectively captures essential information from video sequences of varying durations, and it achieves high compression rates while preserving crucial discriminative information. The proposed framework demonstrates promising performance across various benchmarks, excelling in both long-form video understanding tasks and standard video understanding benchmarks. These results underscore the versatility and efficacy of our approach, particularly in managing the complexities of prolonged video sequences.