Project in Machine Vision
Vuosi | Lukukausi | Päivämäärä | Periodi | Kieli | Vastuuhenkilö |
---|---|---|---|---|---|
2016 | kevät | 20.01-02.03. | 3-3 | Englanti | Mats Sjöberg |
Luennot
Aika | Huone | Luennoija | Päivämäärä |
---|---|---|---|
Ke 12-14 | C220 | Mats Sjöberg | 20.01.2016-02.03.2016 |
Information for international students
This project course will be held in English.
Yleistä
Computer vision is a research discipline motivated by the need of building machines that can see. This has turned out to be a hard task: modern computers can be used for a variety of remarkable feats but they are often outperformed in simple vision tasks by a three-year old child.
Kurssin suorittaminen
The requirements for passing the project are:
- implementation of some computer vision algorithm or algorithms,
- running experiments,
- writing a short scientific report detailing the method, implementation and experiments,
- giving a short (10-15 min) presentation/demo to other students at the end.
If you have already done the accompanying seminar course, and you wish to do your project on the same topic, you may reuse your report from the seminar course, as long as you add a description of the experiments and the implemented algorithm. The full source code must also be submitted separately. in all cases.
The first meeting is on Wednesday, January 20th at 12:15-14:00 in room C220. In this introductory session we will cover the practical arrangements of the course and some examples. After this there will be a weekly help session at the same time and place.
Presentations and/or demos will be given on Wednesday, March 2 in the regular place and time (12:15 in C220).
The deadline for selecting a topic (to be approved by the lecturer) is January 31st.
The final deadline for the report is March 6th.
Kirjallisuus ja materiaali
Material from the introduction lecture:
Useful links:
- Seminar course book: http://computervisionmodels.com/
- Free online books on deep learning: deeplearningbook.org neuralnetworksanddeeplearning.com
- Free online course on Deep Learning
Libraries:
- OpenCV: computer vision - C++, Python, Java
- VLFeat: computer vision - C, Matlab
- dlib: machine learning, image processing - C++, some Python
- scikit-learn Python: NumPy + SciPy + matplotlib
- libsvm: SVM only
- Weka: machine learning with GUI
- Caffe: deep learning, C++ CUDA
- Theano: deep learning, Python
- TensorFlow: deep learning, Python, C++
- Image Processing Toolbox: Matlab
Datasets: