Gavriel Habib — accepted AI-conference papers

2025

CarGait: Cross-Attention based Re-ranking for Gait recognition

Gavriel Habib, Noa Barzilay, Or Shimshi, Rami Ben-Ari, Nir Darshan

ICCV 2025poster

Gait recognition is a computer vision task that identifies individuals based on their walking patterns. Its performance is commonly evaluated by ranking a gallery of candidates and measuring the identification accuracy at Rank-K. Existing models are typically single-staged, searching for the probe's…

Cited by 0 Source ↗PDF ↗Scholar ↗

2025

EffoVPR: Effective Foundation Model Utilization for Visual Place Recognition

Issar Tzachor, Boaz Lerner, Matan Levy, Michael Green, Tal Berkovitz Shalev, Gavriel Habib +5

ICLR 2025poster

The task of Visual Place Recognition (VPR) is to predict the location of a query image from a database of geo-tagged images. Recent studies in VPR have highlighted the significant advantage of employing pre-trained foundation models like DINOv2 for the VPR task. However, these models are often deeme…

Cited by 10 Source ↗PDF ↗Scholar ↗