Subject description - B3B33VIR

Summary of Study | Summary of Branches | All Subject Groups | All Subjects | List of Roles | Explanatory Notes               Instructions
B3B33VIR Robot Vision
Roles:PV Extent of teaching:2P+2L
Department:13133 Language of teaching:CS
Guarantors:Zimmermann K. Completion:Z,ZK
Lecturers:Zimmermann K. Credits:4
Tutors:Azayev T., Vacek P., Zimmermann K. Semester:Z

Anotation:

Course learn to apply the machine learning methods and optimization algorithms on known robotics problems such as metrical or semantic segmentation from RGB-D data or reactive motion control. The focus of the subject lies in teaching deep learning methods. Students employ the elementary knowledge of optimization and linear algebra such as robust solutions of overdetermined systems of nonlinear equations or gradient minimization methods. First 7 labs are devoted to solving elementary problems in PyTorch, second half is devoted to the individual solution of the semester work.

Study targets:

Course learn to apply the machine learning methods and optimization algorithms on known robotics problems such as metrical or semantic mapping from RGB-D data or reactive motion control.

Course outlines:

1 Overview and lecture outline 2 Regression ML/MAP 3 Classification ML/MAP 4 Neural networks, backpropagation 5 Convolution layer + backpropagation 6 Normalization layers (BachNorm, InstanceNorm, ...) + backpropagation 7 Training (SGD, momentum, ...) 8 Architectures of deep neural networks I: detection (yolo), segmentation (DeepLab), classification (ResNet) 9 Architectures of deep neural networks II: pose regression, LIFT 10 Introduction to PyTorch 11 Generative Adversarial Networks, Cascaded Refinement Networks, Style Transfer Networks 12 Reinforcement Learning in Robotics (Imitation Learning, RL, Actor-Critic, applications) 13 Presentation of semestral work

Exercises outline:

During labs, students will work on individual semestral works.

Literature:

Thrun S., Burgard W., Fox D. Probabilistic robotics, MIT Press, 2006 Šonka M., Hlavác V., Boyle R.: Image processing, analysis, and machine vision, Cengage Learning, Toronto, 2015.

Requirements:

Webpage:

https://cw.fel.cvut.cz/wiki/courses/B3B33VIR

Keywords:

Robotics, computer vision, deep learning

Subject is included into these academic programs:

Program Branch Role Recommended semester
BPKYR_2016 Common courses PV 5


Page updated 5.3.2021 17:52:09, semester: Z/2020-1, L/2021-2, L/2020-1, Z/2021-2, Send comments about the content to the Administrators of the Academic Programs Proposal and Realization: I. Halaška (K336), J. Novák (K336)