Research
I'm interested in learning with limited supervision, specifically with the application
to visual recognition.
|
|
LPOSS: Label Propagation Over Patches and Pixels for Open-vocabulary Semantic Segmentation
Vladan Stojnić, Yannis Kalantidis, Jiří Matas, Giorgos Tolias
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025
arxiv /
code /
demo /
This work introduces LPOSS, a training-free method for open-vocabulary semantic segmentation using Vision-Language Models (VLMs). Our approach enhances the initial per-patch predictions of VLMs through label propagation, which jointly optimizes predictions by incorporating patch-to-patch relationships. We address resolution limitations inherent to patch-based encoders by applying label propagation at the pixel level as a refinement step, significantly improving segmentation accuracy near class boundaries.
|
|
ILIAS: Instance-Level Image retrieval At Scale
Giorgos Kordopatis-Zilos, Vladan Stojnić, Anna Manko, Pavel Šuma, Nikolaos-Antonios Ypsilantis, Nikos Efthymiadis, Zakaria Laskar, Jiří Matas, Ondřej Chum, Giorgos Tolias
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025
arxiv /
code /
website /
This work introduces ILIAS, a new test dataset for Instance-Level Image retrieval At Scale. It is designed to evaluate the ability of current and future foundation models and retrieval techniques to recognize particular objects.
|
|
Label Propagation for Zero-shot Classification with Vision-Language Models
Vladan Stojnić, Yannis Kalantidis, Giorgos Tolias
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024
arxiv /
code /
We tailor label propagation to VLMs and zero-shot classification over bi-modal graphs, and propose an efficient way for performing inductive inference with label propagation via a dual solution and through sparsification.
|
|
Training Ensembles With Inliers and Outliers for Semi-Supervised Active Learning
Vladan Stojnić, Zakaria Laskar, Giorgos Tolias
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2024
paper /
arxiv /
code /
We show that simple ingredients such as joint training with inliers and outliers, semi-supervision, and ensembles enable classical active learning methods to achieve state-of-the-art results in the presence of outliers.
|
Research before PhD
During my master's studies, I was doing research focused mainly on remote sensing
applications.
|
|
Self-Supervised Learning of Remote Sensing Scene Representations Using Contrastive Multiview Coding
Vladan Stojnić, Vladimir Risojević
IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2021
arxiv /
code /
We conduct an extensive analysis of the applicability of self-supervised learning in remote sensing image classification and show that self-supervised learning can be easily extended to multispectral images.
|
|
A Method for Detection of Small Moving Objects in UAV Videos
Vladan Stojnić, Vladimir Risojević, Mario Muštra, Vedran Jovanović, Janja Filipi, Nikola Kezić, Zdenka Babić
Remote Sensing, 2021
paper /
code /
We develop a method trained on synthetic data for the detection of tiny moving objects (bees) in the video.
|
|