Word-Anchored Temporal Forgery Localization

2026-03-06

AI Summary

Existing temporal forgery localization methods are hindered by mismatched feature granularity and high computational overhead. This work proposes a word-anchored localization paradigm that reformulates the task as a discrete binary classification problem grounded in natural linguistic boundaries of speech. To align features with task semantics, we introduce a Forensic Feature Realignment (FFR) module and an Asymmetric Counterfeit-Aware (ACA) loss tailored to highlight forgery traces. The resulting approach requires only a lightweight linear classifier and achieves significant performance gains over current methods in both in-domain and cross-domain settings, while simultaneously offering reduced parameter count and improved computational efficiency.

Technology Category

• Natural Language Processing: Language Grounding & Multi-modal NLP
• Machine Learning: Feature Construction/Reformulation
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data

Abstract

Current temporal forgery localization (TFL) approaches typically rely on temporal boundary regression or continuous frame-level anomaly detection paradigms to derive candidate forgery proposals. However, they suffer not only from feature granularity misalignment but also from costly computation. To address these issues, we propose word-anchored temporal forgery localization (WAFL), a novel paradigm that shifts the TFL task from temporal regression and continuous localization to discrete word-level binary classification. Specifically, we first analyze the essence of temporal forgeries and identify the minimum meaningful forgery units, word tokens, and then align data preprocessing with the natural linguistic boundaries of speech. To adapt powerful pre-trained foundation backbones for feature extraction, we introduce the forensic feature realignment (FFR) module, mapping representations from the pre-trained semantic space to a discriminative forensic manifold. This allows subsequent lightweight linear classifiers to efficiently perform binary classification and accomplish the TFL task. Furthermore, to overcome the extreme class imbalance inherent to forgery detection, we design the artifact-centric asymmetric (ACA) loss, which breaks the standard precision-recall trade-off by dynamically suppressing overwhelming authentic gradients while asymmetrically prioritizing subtle forensic artifacts. Extensive experiments demonstrate that WAFL significantly outperforms state-of-the-art approaches in localization performance under both in- and cross-dataset settings, while requiring substantially fewer learnable parameters and operating at high computational efficiency.

Problem

• temporal forgery localization
• feature granularity misalignment
• computation cost
• class imbalance
• forgery detection

Innovation

• word-anchored temporal forgery localization
• forensic feature realignment
• artifact-centric asymmetric loss
• discrete word-level classification
• foundation model adaptation