This app identifies whether an audio file is synthetic or real and uses GradCAM saliency maps to visualize the key audio features that influenced its classification. read more of Audio Deepfake Detection with GradCam Saliency Map Explainability

An information flow-preserving knowledge distillation methodology for model compression. read more of InDistill

Retrieving algorithm component combinations from online (or local) yaml resources. read more of yamlres

Provides a framework for learning text-driven generative paths in pre-trained GANs. read more of ContraCLIP: Interpretable GAN generation driven by pairs of contrasting sentences

Provides a framework for anonymizing faces in public datasets using pre-trained GANs. read more of Attribute-preserving Face Dataset Anonymization via Latent Code Optimization

Provides a framework for discovering non-linear interpretable paths in pre-trained GAN latent spaces. read more of WarpedGANSpace: Finding non-linear RBF paths in GAN latent space

Provides a Neural Face Reenactment framework by leveraging the expressiveness of the StyleGAN2’s style space. read more of StyleMask: Disentangling the Style Space of StyleGAN2 for Neural Face Reenactment

Provides a framework for the problem of Neural Face Reenactment using Generative Adversarial Networks (GANs). read more of HyperReenact: One-Shot Reenactment via Jointly Learning to Refine and Retarget Faces

Provides a video similarity learning approach using self-supervision. read more of Self-Supervised Video Similarity Learning

Provides a framework for addressing the problem of computationally efficient content-based video retrieval in large-scale datasets. read more of DnS: Distill-and-select for efficient and accurate video indexing and retrieval