• People
  • Research
  • Projects
  • Publications
  • Resources
ViCoS Lab

Luka Čehovin Zajc

Assistant Professor
luka.cehovin@fri.uni-lj.si
Twitter: @lukacu
LinkedIn: lukacu
+386 1 479 8252

About me

I am an assistant professor at Visual Cognitive Systems Lab at Faculty of Computer and Information Science, University of Ljubljana. At the faculty I am involved in Multimedia Systems and Robotics and Computer Perception courses.

My research interests are computer vision, machine learning, remote sensing, and human-computer interaction. I am a co-organizer of the VOT Challenge where we work on systematic visual tracker evaluation.

I also have a personal webpage.

Research

Discriminative correlation filter tracking

We explore online learning of target visual models via discriminative correlation filters. The research spans hand-crafted features and optimization techniques for CPU-based tracking as well as deep learning variants with discriminative feature adaptation and online segmentation.

Visual object tracking performance evaluation

One of the problems of visual tracking evaluation is a lack of a consistent evaluation methodology. This is hampering the cross-paper tracker comparison and faster advancement of the field. In our research we investigate different aspects of tracking evaluation. A continuous effort that is a part of our work is also the Visual Object Tracking Challenge (VOT).

Apparent motion patterns

We propose to go beyond pre-recorded benchmarks with post-hoc annotations by presenting an approach that utilizes omnidirectional videos to generate realistic, consistently annotated, short-term tracking scenarios with exactly parameterized motion patterns..

Remote sensing

Remote sensing involves scanning of the earth by satellite or high-flying aircraft and analyzing it. Since the amount of data acquired this way is huge and growing, matchine learning can be used to perform tasks efficently. We are using modern computer vision methods and apply them to different problems in remote sensing.

Local-global visual models in visual tracking

We addresses the problem of tracking objects which undergo rapid and significant appearance changes. We explore coupled-layer visual models that combines the target's global and local appearance.

Learning a hierarchy of parts

We deal with a problem of Multi-class Object Representation and present a framework for learning a hierarchical shape vocabulary capable of representing objects in hierarchical manner using a statistically important compositional shapes. The approach takes simple oriented contour fragments and learns their frequent spatial configurations. These are recursively combined into increasingly more complex and class specific shape compositions, each exerting a high degree of shape variability

ViCoS Eye

ViCoS Eye is an experimental online service that aims to demonstrate a state-of-the-art computer vision object detection and categorization algorithm developed in our laboratory. Web-service is available in a form of a web-page and in a form of an Android application.

Educational robotics

We have used, evaluated and even developed several robotic systems that are used in our teaching activities. We are developing an open-source robot manipulator platform and a low-cost manipulator that the system runs on.

Projects

RoDEO - Robust Deep Learning for Earth Observation​

January 2025 - December 2027
This ARIS funded project investigats the relationship between sensor fusion and self-supervised learning for data-driven Earth Observation. We focus on the role of self-supervised deep learning for sensor fusion from the perspective of different sources with different spatial resolutions and spectral coverage. The project is grounded in a real-world application in the field of hydrology, where the goal is to predict the water level in rivers using satellite and drone imagery.

EOFuseR - Earth Observation with Sensor-Fusion and Representation Learning

January 2025 - April 2026
This ESA funded project investigates the relationship between sensor fusion and self-supervised learning for data-driven Earth Observation. We focus on the role of self-supervised deep learning for sensor fusion from the perspective of different sources with different spatial resolutions and spectral coverage. The project is grounded in a real-world application in the field of hydrology, where the goal is to predict the water level in rivers using satellite and drone imagery.

COMET - Advanced Computer Vision for Understanding Complex Object Motion in Dynamic Environments

January 2025 - December 2027
This project aims to develop a novel motion understanding paradigm, centered on automatically determining the minimal scene understanding required to track one or multiple objects throughout a video. It tackles three core challenges: developing a few-shot object detector capable of identifying all objects in a category based on limited examples, tracking individual objects amid distractors, and extending this to track transformable objects in complex environments.

SMASH - Machine learning for science and humanities postdoctoral program

July 2023 - June 2028
SMASH is an innovative, intersectoral, career-development training program for outstanding postdoctoral researchers, co-funded by the Marie Skłodowska-Curie Actions COFUND program. SMASH is open to researchers around the world who are interested in developing cutting-edge machine learning applications for science and humanities.

MV4.0 - Data-driven framework for development of machine vision solutions

October 2021 - September 2024
The functional objective of the project is to shift the paradigm in the development of machine vision solutions from hand-engineered specific solutions to data-driven learning-based design and development that would enable more general, efficient, flexible and economical development, deployment and maintenance of machine vision systems. The main research goal of this project is to develop novel deep learning methods for iterative, active, robust, weak, self-, unsupervised and few-shot learning that would reduce the amount of needed annotated data.

GAPTRACK - Deep generative appearance modeling in visual tracking

July 2019 - December 2021
The challenge that we address in this project is a robust design of deep generative models, their training and application to a visual tracking scenario. We believe that a generative appearance model of the entire object is a crucial step towards grounding visual object tracking in high-level concepts behind raw pixel values.

Computer Vision

January 2019 - December 2027
Computer vision is becoming a focal problem area of artificial intelligence. On the wings of deep learning it has become very powerful tool for solving various problems involving processing of visual information. In the framework of this programme we are addressing several research questions ranging from visual tracking to visual learning for autonomous robots, with a special emphasis on going beyond supervised deep learning.

TraPri - Tradition meets the future - computer vision and augmented reality for the preservation and promotion of natural and cultural heritage

April 2018 - August 2018
In this student project we have developed an innovative solution based on mobile computer vision and augmented reality, which presents the tradition of viticulture and wine growing in Vipava valley with the technology of the future. We have developed a prototype of an Android mobile application and a content management system that enable efficient and attractive communication of relevant information.

ViAMaRo - Robust computer vision methods for autonomous water surface vehicles

May 2017 - April 2020
The project primary goal is to develop functionalities required for robust autonomous navigation of USVs in uncontrolled environments, primarily relying on the captured visual information. The project focuses on obstacle detection using monocular and stereo systems, development of efficient visual tracking algorithms for marine environments and environment representation through sensor fusion.

HiMoDel - Learning, analysis, and detection of motion in the framework of a hierarchical compositional visual architecture

April 2011 - August 2014
The project primary goal was to develop a holistic approach towards learning, detection and recognition / categorisation of the visual motion and the phenomena derived from it. The project explored the paradigm of learning multi­layer compositional hierarchies.

LeOParts - Learning a large number of visual object categories for content-based retrieval in image and video databases

April 2010 - August 2013
The challenge this project addressed was development of a methodology that would bridge the gap between the computer-centered low-level image features and the high-level human-centered semantic meanings. The methodology explored was hierarchical compositional models, enriched by discriminative information and extended to online learning.

CogX - Cognitive Systems that Self-Understand and Self-Extend

January 2007 - December 2010
The high level aim of this EU FP7 project was to develop a unified theory of self-understanding and self-extension with a convincing instantiation and implementation of this theory in a robot. By self-understanding we mean that the robot has representations of gaps in its knowledge or uncertainty in its beliefs. By self-extension we mean the ability of the robot to extend its own abilities or knowledge by planning learning activities and carrying them out. The project involved six universities and about 30 researchers.

Publications

Resources

vot-toolkit

program
Toolkit to support visual tracking performance evaluation
pythonvottracking

MSKS

program
Schedule and run repeatable taks in Conda environments.
reproduciblecondapython

PixelPipes

library
Infinite data streams for deep learning.
c++pythondeep learningsamplingdata

TraX

library
Tracking Exchange protocol reference implementation
vottrackingcmakepythonmatlabcc++

echolib

library
A simple and efficient interprocess communication (IPC) library for Linux written in C++11.
pythonc++cmakeipc
Faculty of Computer and Information Science

Visual Cognitive Systems Laboratory

University of Ljubljana

Faculty of Computer and Information Science

Večna pot 113
SI-1000 Ljubljana
Slovenia
Tel.: +386 1 479 8245