MTGS: A Novel Framework for Multi-Person Temporal Gaze Following and Social Gaze Prediction. Code : https://github.com/idiap/MTGS
AI & ML interests
Artificial and Cognitive Intelligence
Recent Activity
Papers
Geometric Latent Reasoning Induces Shorter Generations in LLMs
Distilling Conversations: Abstract Compression of Conversational Audio Context for LLM-based ASR
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-RETFound-MAE-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-MAE-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-MAE-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-MAE-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-RETFound-DINOv2-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-DINOv2-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-DINOv2-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-DINOv2-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-DINOv2-Large-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Large-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Large-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Large-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-MAE-Large-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-MAE-Large-Impression
Image Classification • Updated -
Idiap/fm-overspecialization-MAE-Large-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-MAE-Large-CNV
Image Classification • Updated
A multimodal large language model trained specifically for facial image understanding. Project page: https://www.idiap.ch/paper/facellm
We present EdgeFace- a lightweight and efficient face recognition network inspired by the hybrid architecture of EdgeNeXt.
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-RETFound-Green-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-Green-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-Green-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-Green-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-DINOv2-Small-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Small-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Small-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Small-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
Three complementary algorithms, called Langevin, Dispersion, and DisCo, aimed at generating large synthetic face datasets
Face recognition models trained using HyperFace (synthetic face recognition datasets)
Project page: https://www.idiap.ch/paper/hyperface
MTGS: A Novel Framework for Multi-Person Temporal Gaze Following and Social Gaze Prediction. Code : https://github.com/idiap/MTGS
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-RETFound-MAE-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-MAE-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-MAE-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-MAE-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-RETFound-Green-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-Green-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-Green-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-Green-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-RETFound-DINOv2-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-DINOv2-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-DINOv2-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-RETFound-DINOv2-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-DINOv2-Small-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Small-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Small-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Small-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-DINOv2-Large-CNV
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Large-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Large-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-DINOv2-Large-Impression
Image Classification • Updated
Checkpoints from the experiments presented in our paper
Checkpoints from the experiments presented in our paper
-
Idiap/fm-overspecialization-MAE-Large-CNV_ICGA
Image Classification • Updated -
Idiap/fm-overspecialization-MAE-Large-Impression
Image Classification • Updated -
Idiap/fm-overspecialization-MAE-Large-HyperF_Type
Image Classification • Updated -
Idiap/fm-overspecialization-MAE-Large-CNV
Image Classification • Updated
Three complementary algorithms, called Langevin, Dispersion, and DisCo, aimed at generating large synthetic face datasets
A multimodal large language model trained specifically for facial image understanding. Project page: https://www.idiap.ch/paper/facellm
Face recognition models trained using HyperFace (synthetic face recognition datasets)
Project page: https://www.idiap.ch/paper/hyperface
We present EdgeFace- a lightweight and efficient face recognition network inspired by the hybrid architecture of EdgeNeXt.