DeepFense: A Unified, Modular, and Extensible Framework for Robust Deepfake Audio Detection
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917396287062016 |
|---|---|
| author | Kheir, Yassine El Das, Arnab Xiao, Yixuan Wang, Xin Kallel, Feidi Erdogan, Enes Erdem Vu, Ngoc Thang Polzehl, Tim Moeller, Sebastian |
| author_facet | Kheir, Yassine El Das, Arnab Xiao, Yixuan Wang, Xin Kallel, Feidi Erdogan, Enes Erdem Vu, Ngoc Thang Polzehl, Tim Moeller, Sebastian |
| contents | Speech deepfake detection is a well-established research field with different models, datasets, and training strategies. However, the lack of standardized implementations and evaluation protocols limits reproducibility, benchmarking, and comparison across studies. In this work, we present DeepFense, a comprehensive, open-source PyTorch toolkit integrating the latest architectures, loss functions, and augmentation pipelines, alongside over 100 recipes. Using DeepFense, we conducted a large-scale evaluation of more than 400 models. Our findings reveal that while carefully curated training data improves cross-domain generalization, the choice of pre-trained front-end feature extractor dominates overall performance variance. Crucially, we show severe biases in high-performing models regarding audio quality, speaker gender, and language. DeepFense is expected to facilitate real-world deployment with the necessary tools to address equitable training data selection and front-end fine-tuning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_08450 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DeepFense: A Unified, Modular, and Extensible Framework for Robust Deepfake Audio Detection Kheir, Yassine El Das, Arnab Xiao, Yixuan Wang, Xin Kallel, Feidi Erdogan, Enes Erdem Vu, Ngoc Thang Polzehl, Tim Moeller, Sebastian Sound Audio and Speech Processing Speech deepfake detection is a well-established research field with different models, datasets, and training strategies. However, the lack of standardized implementations and evaluation protocols limits reproducibility, benchmarking, and comparison across studies. In this work, we present DeepFense, a comprehensive, open-source PyTorch toolkit integrating the latest architectures, loss functions, and augmentation pipelines, alongside over 100 recipes. Using DeepFense, we conducted a large-scale evaluation of more than 400 models. Our findings reveal that while carefully curated training data improves cross-domain generalization, the choice of pre-trained front-end feature extractor dominates overall performance variance. Crucially, we show severe biases in high-performing models regarding audio quality, speaker gender, and language. DeepFense is expected to facilitate real-world deployment with the necessary tools to address equitable training data selection and front-end fine-tuning. |
| title | DeepFense: A Unified, Modular, and Extensible Framework for Robust Deepfake Audio Detection |
| topic | Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2604.08450 |