Computer Vision

Faculty of Engineering and Computer Science

Version 8.0 of 02/23/2023

11B1121

Computer Vision

• Informatik - Medieninformatik (B.Sc.)
• Informatik - Technische Informatik (B.Sc.)
• Elektrotechnik (B.Sc.)
• Mechatronik (B.Sc.)
• Elektrotechnik im Praxisverbund (B.Sc.)
• Lehramt an berufsbildenden Schulen - Teilstudiengang Informationstechnik (M.Ed.)

2

Seeing and becoming aware of a real-world scene is quite easy for the human visual system, nevertheless how to enable computer doing this? The technical term for the artificial vision of computers is known as computer vision and includes more than capture the real world with photos or video cameras. Indeed, becoming aware and understand a scene is the primary and computationally intensive challenge of computer vision. Without artificial vision application as human-robot-cooperation, face recognition, image-based 3D reconstruction, and automated driving cannot be realized at all.

By participating in this course, you get to know methods, which enabling computers to understand pixel-based captures. Based on each pixel of the cameras recording chips, we first apply so-called classical methods such as template matching, SIFT, SURF, and HOG. Nowadays, artificial intelligence (AI) coupled with image data set from the internet, allow latest computer vision techniques to outperform almost all classical methods. Thus, you will learn the basics of AI-based computer vision like convolutional neural networks recognizing handwriting and generative adversarial networks altering faces, in addition. Equipped with this knowledge from both lecture and practical session, you can evolve any computer or smartphone with a camera into a visual system, capable of seeing and becoming aware of a real-world scene.

1. From pixels to semantic symbols
2. Object detection and recognition
3. Realtime computer vision – vectorized and GPU-boosted algorithms
4. Image classification
5. Image segmentation

Knowledge Broadening
Students taking this course, knowing and understand current techniques of computer vision and are capable of implementing these with the help of existing frameworks. Further, they know the concepts needed for applying neural networks on image data sets.
Knowledge Deepening
Existing knowledge in the areas designing algorithms and video technic will be increased.
Instrumental Skills and Competences
Students are capable of implementing object detection as well as image classification in C++ supported by frameworks such as OpenCV. Besides, they can design and train convolutional neural networks on their own image data sets with frameworks like Keras.
Communicative Skills and Competences
Students are capable of estimating the complexity as well as the applicability of computer vision methods and are capable of discussing these methods with peers. Further, they can give a summary of computer vision techniques to people from outside the field.
Systemic Skills and Competences
Students are capable of creating data sets for AI-based techniques and can estimate the training cost. They are also capable of selecting or combining suitable methods of the classical and AI-based computer vision in order to fit a specific task.

The course is split into lectures and practical sessions. Starting with hand-on-session, the students building first computer vision systems and evaluate their blind spots. Building on these skills, students reimplement algorithms from selected research papers in order to reproduce the results described within the paper.

Sch?ning, Julius

Sch?ning, Julius

5

Dawson-Howe, Kenneth. A practical introduction to computer vision with opencv. John Wiley & Sons, 2014.Kaehler, Adrian, and Gary Bradski. Learning OpenCV 3: computer vision in C++ with the OpenCV library. " O'Reilly Media, Inc.", 2016.Dadhich, Abhinav. Practical Computer Vision: Extract Insightful Information from Images Using TensorFlow, Keras, and OpenCV. Packt Publishing Ltd, 2018.Géron, Aurélien. Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems. O'Reilly Media, 2019.

• Viva Voce
• Placement Report, written

Field Work / Experimental Work

1 Term

Irregular

German and English