Hindi audio-video-Deepfake (HAV-DF)
收藏资源简介:
Hindi audio-video-Deepfake (HAV-DF)数据集是由BML Munjal大学创建的第一个基于印地语的音频-视频深度伪造数据集。该数据集通过人脸交换、唇同步和语音克隆等方法生成,旨在捕捉印地语语音和面部表情的细微差别,为训练和评估印地语环境下的深度伪造检测模型提供坚实基础。数据集的创建过程包括多步骤的深度伪造生成技术,涵盖了从基本修改到高度复杂的伪造。HAV-DF数据集的应用领域主要集中在深度伪造检测和多语言深度伪造识别系统的开发,旨在解决印地语社区中深度伪造带来的隐私、信任和安全问题。
The Hindi Audio-Video Deepfake (HAV-DF) dataset is the first Hindi-focused audio-video deepfake dataset created by BML Munjal University. Generated via methods including face swapping, lip-syncing, and voice cloning, this dataset aims to capture the subtle nuances of Hindi speech and facial expressions, providing a solid foundation for training and evaluating deepfake detection models in Hindi-speaking contexts. The development process of the HAV-DF dataset incorporates multi-stage deepfake generation technologies, ranging from basic modifications to highly sophisticated forgeries. The main application scenarios of the HAV-DF dataset center on the development of deepfake detection and multilingual deepfake recognition systems, with the goal of addressing the privacy, trust, and security issues caused by deepfakes in Hindi-speaking communities.

- 1Hindi audio-video-Deepfake (HAV-DF): A Hindi language-based Audio-video Deepfake DatasetBML Munjal大学工程与技术学院 · 2024年



