|Website:||website containing additional information|
|Period:||period 1 (week 36 through 45, i.e., 3-9-2018 through 9-11-2018; retake week 1)|
|Participants:||up till now 24 subscriptions|
|Schedule:||Official schedule representation can be found in Osiris|
Multimedia retrieval is about the search for and delivery of multimedia documents, such as images, video, audio, biometrics and the combination of these.
This course adopts a signal processing perspective on multimedia and data in general. This perspective allows us to see common aspects in multimedia (e.g., note that an image is a 2D signal) and explore the rich set of techniques, algorithms, and tools signal processing offers us. Where only one or a few signal types will be explored in the assignments and project, the vast majority of techniques, algorithms, and tools can used for other signal types as well. Moreover, this perspective aids the fusion of distinct signals or data types, in a non-conventional way. So, on of the course's main goals is to become aware of the similarities and differences between several types of media (e.g., video, audio, and biosignals) and how to handle this.
In real-world multimedia retrieval and in data mining in general, it becomes increasingly more important to adopt an interdisciplinary stance. This course prepares you on this by discussing core signal processing, modeling via machine learning, and psychological aspects.
Multimedia Retrieval is dominated by emperical computer science, not theoretical computer science. A theoretical foundation is valued; but, its value has to be shown in practice. For example, theoretical complexity can be low; however, emperical complexity can turn out much higher due to a specific hardware architecture. An image processing algorithms can claim an objective optimal retrieval results; however, if the person who assesses the results does not think so, it results will not be evaluated as such.
As a 2nd year MSc-course, the course has the meta goal to prepare its students on the MSc-graduation phase.
Compulsory (exam): The course's handbook: Eidenberger, H. (2012). Handbook of Multimedia Information Retrieval. Vienna, Austria: atpress. ISBN: 978-3-848-22283-4. The book can be ordered at A-Eskwadraat. Alternatively, order the book via Amazon.com or bol.com.
Optional (exam): Oppenheim, A.V. and Schafer, R.W. (2014). Discrete-time signal processing (3rd ed., Pearson New International Edition). Prentice-Hall Signal Processing Series. Harlow, Essex, England: Prentice Education Limited, Inc. ISBN: 978-1-292-02572-8.
Compulsory (project): Olsen, A. (2012). The Tobii I-VT Fixation Filter - Algorithm description. Technical Report. Danderyd, Sweden: Tobii Technology.
|Course form:||The course consists of lectures, practicals, homework, presentations, and a project.|
|Exam form:||The course consists of lectures, practicals/work groups, assignments, and a project, including a presentation and a pitch. Our experience learned us that genuine active participation is needed to pass the course.
Requirement: To pass the course, the weighted average of your assignments, exam, and project has to be at least 5.5 and the grades of the assignments, exam, and the project each have to be at least 4.0.
Retake: Only allowed if the grade of the assignments, exam, and the project is at least a 4.0.
|Minimum effort to qualify for 2nd chance exam:||Om aan de aanvullende toets te mogen meedoen moet de oorspronkelijke uitslag minstens 4 zijn.|
|Description:||This course has objectives to train students in: