Search site

Computer Vision

The activities of the group are presented under two headings: Activity Analysis and Medical Image Analysis.

Activity Analysis

Our research on activity analysis from video has focussed largely on human behaviour. We are interested in all aspects of this problem, including fundamental research on categorisation, tracking, segmentation and motion modelling, through to applied work addressing social and commercial priorities. Some of our recent work, in collaboration with the School's KRR group, is exploring the integration of vision within a broader cognitive framework, including audition, inductive reasoning, and intentionality.

[Carried object detection and tracking picture]

Carried object detection and tracking

Carried object detection is applied using geometric shape properties and tracking is performed using spati-temporal consistency between the object and the person.

[cognito picture]

Activity monitoring and recovery

Learning and predicting activities in an egocentric (First-person viewpoint) setup. The method is applied to the activities of industrial workflows.

[Redvine picture]

Relational description of video scenes

Learning about the activities within a scene, and then about the objects involved in these activities.

[Butterfly picture]

Learning object categories from text description

Can we learn models for fine-grained object categories (butterflies, flowers, etc.) solely from textual descriptions, without using any training images?

[Bicycle Theft Detection]

Bicycle theft detection

Resolving visual uncertainty and ambiguity by linking related events. The method is applied to theft detection in bicycle racks.

[Carried bag picture]

Detecting carried objects

Detecting large objects (e.g. bags) carried by pedestrians depicted in short video sequences.

[Two people playing a table-top game]

Learning to play table-top games

Learn about the objects and patterns of moves used in simple table-top games, and then apply these to play the game.

[Image of intentional map]

Modelling the intentions of pedestrians

Detecting atypical pedestrian trajectories, assuming a simple model of goal-directed navigational behaviour.

Earlier work - but still of interest ...

[Image of intentional map]

Tracking

A generic object tracker, developed initially for tracking vehicles.

[Image of intentional map]

Music via Motion

Create an augmented and interactive audio-visual space.

[Carried bag picture]

Motion segmentation by consensus

Segment moving objects in videos from tracked features

[Carried bag picture]

Watermark extraction

Localise paper-based watermarks using image processing

[Image of intentional map]

Temporal Continuity

Enforcing global spatio-temporal consistency to enhance reliability of moving object tracking and classification.

[Image of intentional map]

Traffic Interactions

Modelling traffic interaction using learnt qualitative spatio-temporal relations and variable length Markov models.

[Image of intentional map]

Electronic stockman's eye

Tracking cows and looking for abnormalities in their movements.

[Image of intentional map]

VLMMs of behaviour

Learning variable length Markov models of human behaviour.

[Image of intentional map]

IMV

Detecting unusual events by modelling simple interactions between people and vehicles.

[Image of intentional map]

Behaviour modelling

Modelling of `object behaviours' using detailed, learnt statistical models.

[Image of intentional map]

Pedestrian tracking

Contour tracking using Active Shape Models.
Source code for the Baumberg tracker.

Medical Image Analysis

We are part of an informal network of groups within the University and Leeds NHS working in the general area of medical imaging, and medical image analysis. These groups include:

Division of Medical Physics
Department of Statistics
Institute of Medical and Biological Engineering
School of Electronic and Electrical Engineering
Leeds Teaching Hospitals NHS Trust

Current and recent projects undertaken in collaboration with this network include:

[Wellmec]

The Wellmec Centre of Excellence

EPSRC/Wellcome Trust funded centre in Medical Engineering

[Image of Histopathology Slide]

Virtual Pathology

Automated analysis of Histopatholgy slides

[Image of Heart MRI]

Cardiac MRI Analysis and Fusion

Automated analysis of multiple Cardiac MRI sequences

[Image of Ultrasound Simulation]

Ultrasound Simulator

Simulation of ultrasound guided needle insertion procedures

[Image of Overview Poster]

Medical Image Analysis and Visualisation Research

A poster giving an overview of medical research undertaken in the group

[Image of virtual suturing]

Simulation of Medical Procedures using Virtual Environments

Recreating a surgical environment using virtual tools

[Image of poster]

Non-rigid multi-modal registration of PET and MRI hand volumes

A model based approach to multi-modal registration

[Image of THz Pulse]

THz

Time frequency analysis techniques in terahertz pulsed imaging. PhD project Sponsored by EPSRC

Collaborators

University of Leeds

Visualisation and Virtual Reality Research Group, School of Computing
Institute of Medical and Biological Engineering
Yorkshire Centre for Health Informatics

Other universities

Department of Informatics, University of Hamburg, Germany
German Research Centre for Artificial Intelligence, Kaiserslautern, Germany
Department of Computer Science, University of Bristol, UK

Other organisations

Leeds Teaching Hospitals NHS Trust
SRI International, Menlo Park, USA