Academic year 2014-15
Face and Gesture Analysis
| Degree: | Code: | Type: |
| Bachelor's Degree in Computer Science | 23104 | Optional subject |
| Bachelor's Degree in Telematics Engineering | 23108 | Optional subject |
| Bachelor's Degree in Audiovisual Systems Engineering | 23112 | Optional subject |
| ECTS credits: | 4 | Workload: | 100 hours | Trimester: | 3rd |
| Department: | Dept. of Information and Communication Technologies |
| Coordinator: | Federico Sukno |
| Teaching staff: | Federico Sukno (theory and seminars/labs) Dmytro Derkach (seminars/labs) |
| Language: | English |
| Timetable: | |
| Building: |
The geometry and dynamics of the human face and body are fundamental sources of information to about individuals. This has been traditionally exploited for machine-based recognition of identity and expressions, but recent trends suggest a much broader scope of applications, including understanding of emotional states, deception clues or even psychological or psychiatric disorders such as autism or depression.
This subject covers the state of the art in face and gesture analysis techniques and provides an overview of the most relevant applications. Covered topics include.
Fundamental knowledge covered by the course includes pattern recognition, probabilistic models, image processing and machine learning. Students will be provided with hands-on exercises to deal with practical face and gesture analysis systems. Hence, basic programming knowledge is advised in at least one of the following: Matlab (recommended) C/C++ or Java.
All classes and materials of the subject (incl. bibliography) will be in English, thus its understanding is required. Students will also be encouraged to present their reports in English, but Spanish or Catalan will be accepted.
| Cross-disciplinary competences | Specific competences |
|---|---|
|
1. Análisis 2. Problem solving 3. Information management 4. Creativity 5. Team work 6. Written communication 7. Work planning and time estimation 8. Practical application of theoretical knowledge 9. Striving for quality 10. Motivation to succeed. |
1. Ability to apply concepts from mathematics, science and engineering. 2. To design and implement experiments 3. To analyze and interpret data or results. 4. Ability to design a system, component or process related to the field of face and gesture analysis. 5. Ability to identify, formulate and solve problems related to face and gesture analysis. 6. Ability to commincate effectively using the technical volabulary of the field in English. 7. Ability to use tools and techniques necessary within face and gesture analysis applications. |
A. Normal Track
|
| Characteristics | Timing | Recoverable | Mark weight |
|---|---|---|---|---|
| Group Project 1 |
Face detection challenge |
Week 4 |
Week 6 |
20% |
| Group Project 2 |
Eigenfaces algorithm |
Week 6 |
Week 8 |
20% |
| Group Project 3 |
Face recognition challenge |
Week 9 |
Week 11 |
30% |
| Written exam |
Final exam (individual) |
End of subject |
July |
30% |
Requirements:
1. All 4 activities listed in the table must be successfully accomplished to pass the subject. In such case, the final mark is a weighted average as specified in the last column of the table.
2. An activity is successfully accomplished if a mark of 5.0 or higher is obtained.
3. Group projects must be performed in teams. Each team will have between 2 and 3 memebers.
5. The 3 teams that achieve the best results in the face recognition challenge can choose avoid the final exam and replace it by a 15-minutes oral presentation of their work.
B. Alternative Project Track
Students who have a clear preference for a specific topic in the subject that is not covered by the organized Group Projects will be allowed to propose a small research project as an alternative to Group Projects 2 and 3. The procedure to access this modality are as follows:
1. Projects must be performed in teams. Each team will have between 2 and 3 memebers.
2. Preliminary discussion with the subject coordinator no later than week 3.
3. Presentation of a detailed proposal no later than week 4.
4. Proposals with sufficient quality will be accepted (up to a maximum of 3 for the whole class). In such case, the team is allowed to continue with the proposed research project which will replace Group Projects 2 and 3.
5. A proposal can be rejected for 2 reasons:
5.a) It has sufficient quality but there are more than 3 proposals. In this case the proposal will be assigned a mark above 5.0 and it will count as Group Activity 2. However, the proposed project cannot proceed and the team has to accomplish Group Project 3 instead.
5.b) It is of insufficient quality. In this case the proposal is not maked and the team returns to the "normal" track, which implies performing Group Projects 2 and 3.
6. Apart from the project itself, teams performing the Alternative Project will have a 15-minute oral presentation of their work (in this case the presentation is mandatory and does not replace the final exam, which has to be taken as well).
The table below summarizes the Alternative Project path:
|
| Characteristics | Timing | Recoverable | Mark weight |
|---|---|---|---|---|
| Group Project 1 |
Face detection challenge |
Week 4 |
Week 6 |
20% |
| Alternative Proposal | Subject to preliminary discussions with subject teachers | Week 4 | Return to Normal Track | 20% |
|
If Alternative Proposal Accepted | ||||
| Alternative Project |
Including written report and oral presentation |
Week 9 |
Week 11 |
30% |
| Written exam |
Final exam (individual) |
End of subject |
July |
30% |
The subject is organized as follows:
Scientific articles (to be provided during the course).