Toward Fine-Grained Speech Inpainting Forensics:A Dataset, Method, and Metric for Multi-Region Tampering Localization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Vu, Tung, Nguyen, Yen, Nguyen, Hai, Pham, Cuong, Tran, Cong
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910233889079296
author Vu, Tung
Nguyen, Yen
Nguyen, Hai
Pham, Cuong
Tran, Cong
author_facet Vu, Tung
Nguyen, Yen
Nguyen, Hai
Pham, Cuong
Tran, Cong
contents Recent advances in voice cloning and text-to-speech synthesis have made partial speech manipulation - where an adversary replaces a few words within an utterance to alter its meaning while preserving the speaker's identity - an increasingly realistic threat. Existing audio deepfake detection benchmarks focus on utterance-level binary classification or single-region tampering, leaving a critical gap in detecting and localizing multiple inpainted segments whose count is unknown a priori. We address this gap with three contributions. First, we introduce MIST (Multiregion Inpainting Speech Tampering), a large-scale multilingual dataset spanning 6 languages with 1-3 independently inpainted word-level segments per utterance, generated via LLM-guided semantic replacement and neural voice cloning, with fake content constituting only 2-7% of each utterance. Second, we propose ISA (Iterative Segment Analysis), a backbone-agnostic framework that performs coarse-to-fine sliding-window classification with gap-tolerant region proposal and boundary refinement to recover all tampered regions without prior knowledge of their count. Third, we define SF1@tau, a segment-level F1 metric based on temporal IoU matching that jointly evaluates region count accuracy and localization precision. Zero-shot evaluation reveals that partial inpainting at word granularity remains unsolved by existing deepfake detectors: utterance-level classifiers trained on fully synthesized speech assign near zero fake probability to MIST utterances where only 2-7% of content is manipulated. ISA consistently outperforms non-iterative baselines in this challenging setting, and the dataset, code, and evaluation toolkit are publicly released.
format Preprint
id arxiv_https___arxiv_org_abs_2605_02223
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Toward Fine-Grained Speech Inpainting Forensics:A Dataset, Method, and Metric for Multi-Region Tampering Localization
Vu, Tung
Nguyen, Yen
Nguyen, Hai
Pham, Cuong
Tran, Cong
Sound
Computer Vision and Pattern Recognition
Recent advances in voice cloning and text-to-speech synthesis have made partial speech manipulation - where an adversary replaces a few words within an utterance to alter its meaning while preserving the speaker's identity - an increasingly realistic threat. Existing audio deepfake detection benchmarks focus on utterance-level binary classification or single-region tampering, leaving a critical gap in detecting and localizing multiple inpainted segments whose count is unknown a priori. We address this gap with three contributions. First, we introduce MIST (Multiregion Inpainting Speech Tampering), a large-scale multilingual dataset spanning 6 languages with 1-3 independently inpainted word-level segments per utterance, generated via LLM-guided semantic replacement and neural voice cloning, with fake content constituting only 2-7% of each utterance. Second, we propose ISA (Iterative Segment Analysis), a backbone-agnostic framework that performs coarse-to-fine sliding-window classification with gap-tolerant region proposal and boundary refinement to recover all tampered regions without prior knowledge of their count. Third, we define SF1@tau, a segment-level F1 metric based on temporal IoU matching that jointly evaluates region count accuracy and localization precision. Zero-shot evaluation reveals that partial inpainting at word granularity remains unsolved by existing deepfake detectors: utterance-level classifiers trained on fully synthesized speech assign near zero fake probability to MIST utterances where only 2-7% of content is manipulated. ISA consistently outperforms non-iterative baselines in this challenging setting, and the dataset, code, and evaluation toolkit are publicly released.
title Toward Fine-Grained Speech Inpainting Forensics:A Dataset, Method, and Metric for Multi-Region Tampering Localization
topic Sound
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.02223