4 papers
A Simple Framework for Open-Vocabulary Zero-Shot Segmentation
Thomas Stegmüller, Tim Lebailly, Nikola Dukic +3
Zero-shot classification capabilities naturally arise in models trained within a vision-language contrastive framework. Despite their classification prowess, these models struggle…
Synthetic Captions for Open-Vocabulary Zero-Shot Segmentation
Tim Lebailly, Vijay Veerabadran, Satwik Kottur +2
Generative vision-language models (VLMs) exhibit strong high-level image understanding but lack spatially dense alignment between vision and language modalities, as our findings in…
Collapse-Proof Non-Contrastive Self-Supervised Learning
Emanuele Sansone, Tim Lebailly, Tinne Tuytelaars
We present a principled and simplified design of the projector and loss function for non-contrastive self-supervised learning based on hyperdimensional computing. We theoretically…
Object-Centric Pretraining via Target Encoder Bootstrapping
Nikola ÄukiÄ, Tim Lebailly, Tinne Tuytelaars
Object-centric representation learning has recently been successfully applied to real-world datasets. This success can be attributed to pretrained non-object-centric foundation mod…