Computer Vision II
Prof. Thomas BroxComputer Vision is a very active research field with many practical applications, for instance in quality control, robotics, driver assistance systems, and many more. The ultimate goal of computer vision is to imitate the great capabilies of the human visual system, allowing the computer not only to record images but also to interpret them. Research is still far from this goal, but significant progress has been made in recent years. This course continues the course Computer Vision I from the last semester. It will deal specifically with 3D reconstruction from multiple views (stereo, structure-from-motion) and recognition. The exercises will consist of programming assignments (in C/C++) in the first half of the semester, where students will learn to implement some selected techniques presented in class. The exercises in the second half of the semester will consist of one larger project assignment in the scope of object recognition. It will be concluded by a short written summary of the work corresponding to a scientific paper. Apart from gathering technical knowledge, students are supposed to train their independent thinking skills and to learn using concept knowledge in order to solve new problems.
This course will not be offered in summer 2015. Exams will be offered to those students who already finished the project last year or those who need only 4 ECTS points. Please check carefully if a course with 4 ECTS points is compatible with your study plan.
|
|
Slides:
Class 15 (8.5.): 3D geometry and camera calibrationClass 16 (15.5.): Stereo reconstruction
Class 17 (22.5.): Disparity estimation
Class 18: Point matching and point tracking to be skipped because of too many public holidays. Please watch the recordings
Class 19 (5.6.): Pose estimation and tracking
Class 20 (26.6.): Structure from motion (by Benjamin Ummenhofer)
Class 21 (3.7.): Dense volumetric multiview reconstruction
Class 22 (10.7.): Instance recognition
Class 23 (17.7.): Object class recognition
Class 24 (24.7.): Part based object detectors
Class 25 (31.7.): Object segmentation with shape priors
Recordings (from last year):
About 100MB each.Class 15: 3D geometry and camera calibration
Class 16: Stereo reconstruction
Class 17: Disparity estimation
Class 18: Point matching and point tracking
Class 19: Pose estimation and tracking
Class 20: Structure from motion
Class 21: Dense volumetric multiview reconstruction
Class 22: Instance recognition
Class 23: Object class recognition
Class 24: Part based object detectors
Class 25: Object segmentation with shape priors
Exercise material:
Exercise 1 (8.5.)Exercise 2 (15.5.)
Exercise 3 (22.5.)
Discussion and assignment of project (5.6.)
Project work (remaining dates)