AKCIT-Deepfake/BRSpeech-DF

BRSpeech-DF是一个葡萄牙语深度伪造语音检测数据集，包含459,000个音频样本，包括真实人类语音和通过多种零样本文本到语音(TTS)模型生成的合成语音。该数据集旨在促进更加健壮、包容和多语言的音频深度伪造检测系统的发展，尤其是对于葡萄牙语等代表性不足的语言。

BRSpeech-DF is the first publicly available dataset for deepfake speech detection in Portuguese, containing 459,000 audio samples including both real human speech and synthetic speech generated using multiple zero-shot text-to-speech (TTS) models. This dataset aims to promote the development of more robust, inclusive, and multilingual audio deepfake detection systems, especially for underrepresented languages such as Portuguese.