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

arXiv: https://arxiv.org/abs/2605.02223 · HTML: https://mlsanity.com/paper/2605.02223/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 23 of its 37 references that are in the map.
- First posted on arXiv 4 May 2026.

## Builds on (most cited first, 23 of 23)

- [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://mlsanity.com/paper/2006.11477/) (2020)
- [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://mlsanity.com/paper/2106.07447/) (2021)
- [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://mlsanity.com/paper/2110.13900/) (2021)
- [ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification](https://mlsanity.com/paper/2005.07143/) (2020)
- [AudioLM: a Language Modeling Approach to Audio Generation](https://mlsanity.com/paper/2209.03143/) (2022)
- [MLS: A Large-Scale Multilingual Dataset for Speech Research](https://mlsanity.com/paper/2012.03411/) (2020)
- [CosyVoice: A Scalable Multilingual Zero-shot Text-to-speech Synthesizer based on Supervised Semantic Tokens](https://mlsanity.com/paper/2407.05407/) (2024)
- [AASIST: Audio Anti-Spoofing using Integrated Spectro-Temporal Graph Attention Networks](https://mlsanity.com/paper/2110.01200/) (2021)
- [End-to-end anti-spoofing with RawNet2](https://mlsanity.com/paper/2011.01108/) (2020)
- [ASVspoof 2021: accelerating progress in spoofed and deepfake speech detection](https://mlsanity.com/paper/2109.00537/) (2021)
- [CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-training](https://mlsanity.com/paper/2505.17589/) (2025)
- [ASVspoof 2021: Towards Spoofed and Deepfake Speech Detection in the Wild](https://mlsanity.com/paper/2210.02437/) (2022)
- [Automatic speaker verification spoofing and deepfake detection using wav2vec 2.0 and data augmentation](https://mlsanity.com/paper/2202.12233/) (2022)
- [ADD 2022: the First Audio Deep Synthesis Detection Challenge](https://mlsanity.com/paper/2202.08433/) (2022)
- [AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset](https://mlsanity.com/paper/2311.15308/) (2023)
- [ASVspoof 2019: spoofing countermeasures for the detection of synthesized, converted and replayed speech](https://mlsanity.com/paper/2102.05889/) (2021)
- [Audio Deepfake Detection: A Survey](https://mlsanity.com/paper/2308.14970/) (2023)
- [The PartialSpoof Database and Countermeasures for the Detection of Short Fake Speech Segments Embedded in an Utterance](https://mlsanity.com/paper/2204.05177/) (2022)
- [Half-Truth: A Partially Fake Audio Detection Dataset](https://mlsanity.com/paper/2104.03617/) (2021)
- [An Initial Investigation for Detecting Partially Spoofed Audio](https://mlsanity.com/paper/2104.02518/) (2021)
- [LlamaPartialSpoof: An LLM-Driven Fake Speech Dataset Simulating Disinformation Generation](https://mlsanity.com/paper/2409.14743/) (2024)
- [LEMAS: Large A 150K-Hour Large-scale Extensible Multilingual Audio Suite with Generative Speech Models](https://mlsanity.com/paper/2601.04233/) (2026)
- [Analyzing the Impact of Splicing Artifacts in Partially Fake Speech Signals](https://mlsanity.com/paper/2408.13784/) (2024)

## Abstract (arXiv)

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.