About

I am currently a DPhil (PhD) student at University of Oxford, AIMS CDT, advised by Prof A. Vedaldi, Dr C. Rupprecht, and Dr I. Laina at VGG. I am broadly interested in unsupervised scene understanding, with emphasis on objects.

In previous life, I worked as MLE for OakNorth and Bloomberg. I graduated with MEng in Computer Science from University of Cambridge, supervised by Prof Liò.

I am always happy to discuss research, so feel free to reach out!

Interests

  • Unsupervised methods
  • Generative models
  • Computer vision
  • Scene understanding

Education

  • DPhil, (ongoing)

    University of Oxford

  • MEng in Computer Science, 2017

    University of Cambridge

  • BA in Computer Science, 2017

    University of Cambridge

Publications

Unsupervised Multi-object Segmentation by Predicting Probable Motion Patterns

We propose a new approach to learn to segment multiple image objects without manual supervision. The method can extract objects form …
Laurynas Karazija*, Subhabrata Choudhury*, Iro Laina, Christian Rupprecht, Andrea Vedaldi
PDF Code Project Video
Unsupervised Multi-object Segmentation by Predicting Probable Motion Patterns

Guess What Moves: Unsupervised Video and Image Segmentation by Anticipating Motion

Motion, measured via optical flow, provides a powerful cue to discover and learn objects in images and videos. However, compared to …
Subhabrata Choudhury*, Laurynas Karazija*, Iro Laina, Andrea Vedaldi, Christian Rupprecht
PDF Code Project Video
Guess What Moves: Unsupervised Video and Image Segmentation by Anticipating Motion

ClevrTex: A Texture-Rich Benchmark for Unsupervised Multi-Object Segmentation

There has been a recent surge in methods that aim to decompose and segment scenes into multiple objects in an unsupervised manner, …
Laurynas Karazija, Iro Laina, Christian Rupprecht
PDF Code Project Poster Video
ClevrTex: A Texture-Rich Benchmark for Unsupervised Multi-Object Segmentation

Automatic Inference of Cross-modal Connection Topologies for X-CNNs

This paper introduces a way to learn cross-modal convolutional neural network (X-CNN) architectures from a base convolutional network …
Laurynas Karazija, Petar Veličković, Pietro Liò
PDF Code Slides DOI

Cross-modal Recurrent Models for Weight Objective Prediction from Multimodal Time-series Data

We analyse multimodal time-series data corresponding to weight, sleep and steps measurements. We focus on predicting whether a user …
Petar Veličković, Laurynas Karazija, Nicholas D Lane, Sourav Bhattacharya, Edgar Liberis, Pietro Liò, Angela Chieh, Otmane Bellahsen, Matthieu Vegreville
PDF DOI