| TSBB17 |
Visual Object Recognition and Detection, 6 ECTS credits.
/Visuell detektion och igenkänning/
For:
D
IT
U
Y
|
| |
Prel. scheduled
hours: 48
Rec. self-study hours: 112
|
| |
Area of Education: Technology
Main field of studies: Computer Engineeirng
|
| |
Advancement level
(G1, G2, A): A
|
|
Aim:
After the course the students should be able to:
- identify basic terminologies, theories and methods for recognition and detection of objects in images
- understand current approaches for object recognition and detection, to actively analyse their strengths and weaknesses
- develop, experimentally evaluate different recognition/detection algorithms and summarize the results
- select appropriate methods for automatic training of recognition and detection systems
- understand basic theories of how the brain processes visual information to perform object recognition and detection tasks
|
|
Prerequisites: (valid for students admitted to programmes within which the course is offered)
Basic image processing: thresholding, segmentation, edge detection from for example Signals, Information and Images or Digital Image Processing. Use of Matlab. Probability Theory and Statistics.
Note: Admission requirements for non-programme students usually also include admission requirements for the programme and threshhold requirements for progression within the programme, or corresponding.
|
|
Organisation:
The course consists of two parts that are presented in parallel. One part is more theoretical and is based on a larger number of lectures that present and illustrate basic methods for object recognition and detection. This part concludes with a written examination. The other part is more practical and begins with an introduction to two projects: one in the area of object recognition and the other in the area of object detection. The course participants are divided into small groups, and each group carries out both these applied projects, which shall demonstrate a number of methods presented in the
theoretical part of the course. The results from each project group are presented orally at seminars and are documented in reports. Guidance for the projects is only given during the course semester. Each project is concluded by an analysis and reflection of the project work.
|
|
Course contents:
Invariant local features and feature extraction in digital images, bag-of-features framework, principles of object recognition and detection, local spatial constraints, introduction to convolutional neural networks, support vector machines learning, shape descriptors and matching, part-based models for recognition, the role of context in recognition, overview of object recognition in biological systems and deep features.
|
|
Course literature:
There is no required textbook for the course. Material will be handed out or made available on the course web page. We will obtain most of our content from the papers we read. Further, Steve Palmer's Vision Science, Richard Szeliski's Computer Vision: Algorithms and Applications, and Forsyth and Ponce's Computer vision: A Modern Approach have useful source material.
|
|
Examination: |
|
Written examination
Oral and written presentation of project assignment
|
3 ECTS
3 ECTS
|
| |
|
The course has a written examination that includes the theoretical and method describing part of the course. The project assignments consist of implementation, report writing, and an oral presentation. |
Course language is English.
Department offering the course: ISY.
Director of Studies: Klas Nordberg
Examiner: Fahad Khan
Course Syllabus in Swedish
|