OBJECT2D3D - Object (category) Detection, Localisation, Registration and Recognition

Object 01 Object 02 Object 03


The goal of this project is to solve some of the most fundamental problems in computer vision - detection, localisation and recognition of real 3-D objects in 2-D (or lately 2.X-D) images. The main focus is set to general methods which are automatic and perform accurately and reliably in many application areas. Methods should tolerate normal distortions in images (e.g. projective transformation, illumination, occlusion and noise),

This project is mainly carried out using academic funding from several sources and the methods are developed gradually in PhD theses during the course of work.

The current main objectives: Approaches currently under investigation:


The project is currently divided to the following PhD tasks:
  1. EVEX - Object Evidence Extraction
  2. SPATMODEL - Spatial Constellation Models of Local Descriptors
The following projects benefit from the results of this project:
  1. FACEDETECT - Face Detection


The latest results of a very accurate object localisation will be presented in the ICCV2007 workshop!

Contact information

Joni Kämäräinen E-mail WWW Project leader, professor
Alexander Drobchenko E-mail WWW Post-graduate researcher
Jarmo Ilonen E-mail WWW Post-graduate researcher
Pekka Paalanen E-mail WWW Post-graduate researcher


Source code

  1. -

Data sets

  1. -

Links to other resources

Research groups working on general object detection and recognition
CalTech Vision Group (prof Perona)
Univ. Oxford - Robotics Group (Visual Geometry Group) (Prof Zisserman)
Princeton Vision Lab (prof Li)
Univ. British Columbia - Lab for Computational Intelligence (prof Lowe)
Cornell Univ. - Computer Vision Group (prof Huttenlocher)
Data sets
Prof Li's group's 3D object category data set from ICCV2007


Do not hesitate to to contact authors (e-mail, etc.) in order to retrieve copies or reprints of the following publications.