MLT-Dedup: Efficient Large-Scale Online Video Deduplication via Multi-Level Representations and Spatial-Temporal Matching

2026-06-10

AI Summary

This work addresses the challenge of degraded user experience and increased storage and bandwidth costs caused by massive near-duplicate videos on online platforms, where existing methods struggle to balance recall and efficiency under limited indexing budgets. The authors propose MLT-Dedup, a novel framework that integrates multi-granularity video encoding (ML-VE) with a difference-aware spatiotemporal similarity module (DiF-SiM). It efficiently retrieves candidates using sparse segment-level embeddings and refines matching at the frame level to achieve high-precision deduplication. Evaluated on a real-world large-scale platform, the method reduces online duplication rates by 91% at 90% precision while increasing indexing capacity fivefold, substantially expanding the coverage of duplicate detection.

Abstract

The explosive growth of user-generated video content on online platforms is accompanied by the emergence of numerous near-duplicate videos--videos that are identical or highly similar but differ by partial edits. These duplicates degrade user experience and increase storage and bandwidth costs, making large-scale video deduplication a critical task. Existing video deduplication frameworks face a fundamental challenge in retrieving sufficient high-quality candidates under a limited index budget, as well as trade-offs between efficiency and precision. To address these issues, we propose MLT-Dedup, an efficient large-scale online video deduplication framework with Multi-Level representations and spatial-Temporal matching. Our approach employs a Multi-Level Video Encoder (ML-VE) to extract both fine-grained frame-level and sparse clip-level embeddings: sparse embeddings support efficient candidate retrieval, while fine-grained embeddings are loaded for precise pairwise matching. During matching, we introduce DiF-SiM, a Differential Feature-enhanced Similarity Module capable of locating duplicated temporal segments and providing reliable similarity evidence to support policy-driven deduplication decisions. Extensive experiments on a real-world large-scale platform demonstrate that MLT-Dedup reduces online repetition rates by 91% at 90% precision. Furthermore, our sparse retrieval design achieves a 5x increase in indexing capacity, enabling broader candidate coverage in real-world deployment.

Problem

• video deduplication
• near-duplicate videos
• large-scale
• online platforms
• index budget

Innovation

• Multi-Level Representations
• Spatial-Temporal Matching
• Video Deduplication
• Sparse Retrieval
• Differential Feature-enhanced Similarity