IDiff-Face

该数据集提供了一个框架，该框架利用基于身份上下文的扩散模型来生成真实图像的各种变体。此外，它还允许使用合成人脸嵌入生成新颖的合成身份。其任务领域为人脸识别。

This dataset presents a framework that leverages identity-context-based diffusion models to generate diverse variants of real-world images. Furthermore, it supports the generation of novel synthetic identities via synthetic face embeddings. Its task domain is face recognition.