Large scale visual recognition of object instances and categories

Monday, September 23, 2013 to Wednesday, September 25, 2013

Infotech Oulu Doctoral Program

Large scale visual recognition of object instances and categories

Lecturer: Dr. Andrea Vedaldi, University of Oxford, UK

Date: September, 23-25, 2013

Time & room:

Monday, September 23, 9:30, IT113
Tuesday, September 24,  PR102
Wednesday, September 25,  PR119
Exercises TS135

 

Course material: http://www.ee.oulu.fi/research/imag/courses/Vedaldi/

 

 

The goal of this school is to introduce a number of state-of-the-art fundamental techniques in image understanding as well as to demonstrate the use of open source software to implement them in applications.

Theoretical aspects that will be covered include image representations suitable for registration, object instance, and object category matching (including regions of interest, local descriptors, descriptor metrics, quantisation, indexing, and histogramming) as well as machine learning techniques to train models for given object types (linear and non-linear large scale support vector machines and related kernel representations and optimisation methods).

Alternating with the theoretical sessions, in a series of guided experiments the students will explore how such ideas can be implemented in software by using MATLAB and open source libraries such as VLFeat.

Andrea Vedaldi is University Lecturer in Engineering Science at the University of Oxford since 2012. His research interests include the automatic interpretation of image, machine learning and large scale optimisation. He is author of more than thirty papers in major computer vision and machine learning conferences and journals, as well as leading author of the VLFeat computer vision library. From 2008 to 2012 he was postdoctoral researcher and junior research fellow at the University of Oxford, supported by the Glasstone Research Fellowship in Science and the New College W. W. Spooner Fellowships. He is the recipient of the PhD and MSc degrees in Computer Science from the University of California at Los Angeles in 2008 and 2005 respectively (outstanding PhD and MSc thesis awards), and of the BSc degree in Information Engineering by the University of Padua in 2003.

Day 1: Matching and recognition of object instances

- Introduction  9.30 - 9.35
- The image matching problem  9.35 - 9.50
- Local features: detectors and descriptors 9.50 - 10.20
- Matching and recognition using local features 10.20 - 10.50
- Dataset and evaluation  10.50 - 11.05
Coffee break (no free coffee, unfortunately   11.05 - 11.35
- Efficient visual search: hashing and indexing 11.35 - 12.30
Lunch 12.30 - 14.30
- Large scale retrieval and applications 14.30 - 15.30
- Practical session on object instance recognition 15.30 - 18.00
- Wrap-up: summary and questions + open research  18.00 - 18.30

Day 2: Learning and recognition of object categories

- Object categories and intra-class variability 9.30 - 9.40
- Machine learning fundamentals 9.40 - 10.30
- Bag of visual words models for image classification 10.30 - 11.00
Coffee break  (no free coffee, unfortunately  11.00 - 11.30
- Advanced image representations 11.30 - 11.40
- Dataset and evaluation 11.40 - 12.00
- Optimization methods for large scale linear learning 12.00 - 12.30
Lunch 12.30 - 14.30
- Large scale non-linear learning with feature maps 14.30 - 15.30
- Practical session on object category recognition 15.30 - 18.00
- Wrap-up: summary and questions + open research 18.00 -18.30

Day 3:  Detecting object categories and advanced learning
- From recognizing to detecting objects 9.30 - 9.40
- HOG detectors  9.40 - 10.00
- Advanced HOG-based representation 10.00 - 10.20
- Advanced object detection 10.20 - 10.50
- Dataset and evaluation 10.50 - 11.00
Coffee break (no free coffee, unfortunately  11.00 - 11.30
- Learning to compare & compress 11.30 - 12.30
Lunch 12.30 - 14.30
- Structured-output support vector machines 14.30 - 15.45
- Wrap-up: Recapitulation and current research challenges 15.45 - 16.00

More information: Esa Rahtu

Add to calendar

Back to events

Last updated: 28.3.2014