8 papers
Retrieval-Augmented Gaussian Avatars: Improving Expression Generalization
Matan Levy, Gavriel Habib, Issar Tzachor +5
Template-free animatable head avatars can achieve high visual fidelity by learning expression-dependent facial deformation directly from a subject's capture, avoiding parametric fa…
OmnimatteZero: Fast Training-free Omnimatte with Pre-trained Video Diffusion Models
Dvir Samuel, Matan Levy, Nir Darshan +2
In Omnimatte, one aims to decompose a given video into semantically meaningful layers, including the background and individual objects along with their associated effects, such as…
Find your Needle: Small Object Image Retrieval via Multi-Object Attention Optimization
Michael Green, Matan Levy, Issar Tzachor +3
We address the challenge of Small Object Image Retrieval (SoIR), where the goal is to retrieve images containing a specific small object, in a cluttered scene. The key challenge in…
CarGait: Cross-Attention based Re-ranking for Gait recognition
Gavriel Habib, Noa Barzilay, Or Shimshi +2
Gait recognition is a computer vision task that identifies individuals based on their walking patterns. Gait recognition performance is commonly evaluated by ranking a gallery of c…
Lightning-Fast Image Inversion and Editing for Text-to-Image Diffusion Models
Dvir Samuel, Barak Meiri, Haggai Maron +5
Diffusion inversion is the problem of taking an image and a text prompt that describes it and finding a noise latent that would generate the exact same image. Most current determin…
Task-Specific Adaptation with Restricted Model Access
Matan Levy, Rami Ben-Ari, Dvir Samuel +2
The emergence of foundational models has greatly improved performance across various downstream tasks, with fine-tuning often yielding even better results. However, existing fine-t…