EVEX - Object Evidence Extraction

Evidences = image features = keypoints = interest points
Local descriptors = interest point descriptors

Goals

The goal of this project is to study existing and develop new methods for accurate detection of local image features aka object evidences. The detection of local image features is needed in the recently emerged object detection, localisation and recognition approach where image features (local descriptors) are separately detected and then "hypotheses" of object location and pose are searched based on proper spatial configuration (constellation model) of the features.

This project is a sub-project of more general project, OBJECT - Object Detection and Recognition, and the results will be utilized in the main project.

The current main objectives are:

State-of-the-art method for extraction of local image features (supervised approach)

Approaches currently under investigation:

Multiresolution Gabor feature based local descriptor
Descriptor ranking and classification based on the Gaussian mixture model

Sub-projects

This project is currently carried out in the following sub-projects:

SIMPLEGABOR - Simple Gabor features
GMMBAYES - Statistical ranking of features

Old version of the evidence extraction method was applied in the following project:

FACEDETECT - Face Detection

News

The proposed state-of-the-art image feature extraction method was accepted to be published in IEEE-TIP and the results will be published in the PhD thesis of Jarmo Ilonen.

Contact information

Joni Kämäräinen	E-mail	WWW	Project leader, lecturer
Jarmo Ilonen	E-mail	WWW	Post-graduate researcher
Pekka Paalanen	E-mail	WWW	Post-graduate researcher
Heikki Kälviäinen	E-mail	WWW	Professor

Download

Source code

Data sets

The following publicly available data sets have been used in our experiments:

The Extended M2VTS Database (XM2VTS) (only colour still images used) (by CVSSP, University of Surrey, UK).
- Training face landmarks: XM2TrClC1/
- Evaluation face landmarks: not used
- Testing face landmarks (d_eye used): XM2TeImC1/
- XM2VTS/non-frontal (mpeg7) ground truth annotated by Joni:
  - Training set: xm2vts_nonfrontal_training_www.txt
  - Testing set: xm2vts_nonfrontal_testing_www.txt
  - Ground truth files: xm2vts_nonfrontal_joni.gt.tar.gz
- For developing spatial constellation search methods you may find useful our 100 best landmark candidate of each type (automatically extracted as reported in our IEEE-TIP2007 article):
  - XM2VTS/frontal: evex_xm2vts_frontal_100best.tar.gz
  - XM2VTS/non-frontal: evex_xm2vts_nonfrontal_100best.tar.gz
The Banca Database (by CVSSP, University of Surrey, UK).

Links to other resources

Researchers working on local descriptors or other similar methods
Dr Cordelia Schmid (INRIA)
Prof David Lowe (University of British Columbia)
Dr Timor Kadir (University of Oxford)

Documentation

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

Articles in international scientific journals with referee practice
1. Ville Kyrki, Joni-Kristian Kämäräinen, Heikki Kälviäinen, Pattern Recognition Letters, Simple Gabor Feature Space for Invariant Object Recognition, (2004) pp. 311--318, No. 3 Vol. 25
2. Miroslav Hamouz, Josef Kittler, Joni-Kristian Kämäräinen, Pekka Paalanen, Heikki Kälviäinen, Jiri Matas, IEEE Transactions on Pattern Analysis and Machine Intelligence, Feature-Based Affine-Invariant Localization of Faces, (2005) pp. 1490-1495, No. 9 Vol. 27
3. Joni-Kristian Kämäräinen, Ville Kyrki, Heikki Kälviäinen, IEEE Transactions on Image Processing, Invariance Properties of Gabor Filter Based Features - Overview and Applications, (2006) pp. 1088-1099, No. 5 Vol. 15
4. Pekka Paalanen, Joni-Kristian Kämäräinen, Jarmo Ilonen, Heikki Kälviäinen, Pattern Recognition, Feature Representation and Discrimination Based on Gaussian Mixture Model Probability Densities - Practices and Algorithms, (2006) pp. 1346-1358, No. 7 Vol. 39
5. Jarmo Ilonen, Joni-Kristian Kämäräinen, Pekka Paalanen, Miroslav Hamouz, Josef Kittler, Heikki Kälviäinen, IEEE Transactions on Image Processing, Image feature localization by multiple hypothesis testing of Gabor features, (2008) pp. 311-325, No. 3 Vol. 17
Articles in international compilation works and in international scientific conference proceedings with referee practice
1. Joni-Kristian Kämäräinen, Ville Kyrki, Miroslav Hamouz, Josef Kittler, Heikki Kälviäinen, Proc. of the IAPR Workshop on Machine Vision Applications, Invariant Gabor Features for Face Evidence Extraction, (2002) pp. 228-231,
2. Miroslav Hamouz, Josef Kittler, Joni-Kristian Kämäräinen, Heikki Kälviäinen, Proc. of the 4th Int. Conf. on the Audio- and Video-Based Biometric Person Authentication, Affine Invariant Localization of Faces in Verification Systems, (2003) pp. 276-284,
3. Miroslav Hamouz, Josef Kittler, Joni-Kristian Kämäräinen, Heikki Kälviäinen, Proc. of the 4th Int. Conf. on the Audio- and Video-Based Biometric Person Authentication (AVBPA2003), Hypotheses-Driven Affine Invariant Localization of Faces in Verification Systems, (2003) pp. 276-284,
4. Ville Kyrki, Joni-Kristian Kämäräinen, Heikki Kälviäinen, Proceedings of the 5th International Conference on Advanced Conce pts for Intelligent Vision Systems Theory and Applications, Simple Gabor Feature Space for Invariant Object Recognition, (2003) pp. 46--51,
5. Joni-Kristian Kämäräinen, Jarmo Ilonen, Pekka Paalanen, Miroslav Hamouz, Heikki Kälviäinen, Josef Kittler, Proc. of the 14th Scandinavian Conf. of Image Analysis, Object Evidence Extraction Using Simple Gabor Features and Statistical Ranking, (2005) pp. 119-129,
6. Jarmo Ilonen, Joni-Kristian Kämäräinen, Proc. of the Int. Joint Conf. on Neural Networks (IJCNN), Object categorization using self-organization over visual appearance, (2006) pp. 4549-4553,
7. Jarmo Ilonen, Joni-Kristian Kämäräinen, Heikki Kälviäinen, Proc. of the 14th International Conference on Image Analysis and Processing (ICIAP2007), Fast extraction of multi-resolution Gabor features, (2007) pp. 481-486,
8. Joni-Kristian Kämäräinen, Miroslav Hamouz, Josef Kittler, Pekka Paalanen, Jarmo Ilonen, Alexander Drobchenko, Proc. of the ICCV 2007 Workshop on Non-Rigid Registration and Tracking Through Learning (NRTL2007), Object Localisation Using Generative Probability Model for Spatial Constellation and Local Image Features, (2007)
Scientific monographs
1. Joni-Kristian Kämäräinen, Ph.D thesis, Feature Extraction Using Gabor Filters, (2003) pdf (2.47 MB)
2. Jarmo Ilonen, Ph.D thesis, Supervised local image feature detection, (2007) pdf (10.26 MB)
Other Scientific Publications
1. Jarmo Ilonen, Joni-Kristian Kämäräinen, Heikki Kälviäinen, Efficient Computation of Gabor Features, (2005) No. 100 pdf (1.06 MB)
2. Pekka Paalanen, Joni-Kristian Kämäräinen, Jarmo Ilonen, Heikki Kälviäinen, Feature Representation and Discrimination Based on Gaussian Mixture Model Probability Densities - Practices and Algorithms, (2005) No. 95 pdf (0.96 MB)
3. Ville Kyrki, Joni-Kristian Kämäräinen, Simple Gabor Feature Space for Invariant Object Recognition, (2003) No. 83