Course Details

Return to Classes
Course Information
SemesterCourse Unit CodeCourse Unit TitleT+P+LCreditNumber of ECTS CreditsLast Updated Date
1ECE532DEEP LEARNING FOR COMPUTER VISION3+0+037,514.05.2025

 
Course Details
Language of Instruction English
Level of Course Unit Doctorate's Degree
Department / Program ELECTRICAL AND COMPUTER ENGINEERING
Type of Program Formal Education
Type of Course Unit Elective
Course Delivery Method Face To Face
Objectives of the Course The aim of this course is to teach the application of deep learning techniques to computer vision problems. It is aimed to provide students with theoretical and practical knowledge of how deep neural networks, especially convolutional neural networks (CNN), which are built on the basic concepts in image processing and computer vision, are used in tasks such as segmentation, object detection, and pose estimation.
Course Content The course systematically covers the fundamentals of deep learning and its applications in computer vision. Topics include artificial neural networks, advanced CNN architectures (ResNet, DenseNet), object detection (YOLO, Faster R-CNN), semantic and attributive segmentation (U-Net, DeepLab), pose estimation, image-to-image translation (GAN-based models), and visual attention mechanisms (Transformers, Vision Transformers). In addition, the data requirements of these methods, training processes, and approaches such as transfer learning are discussed.
Course Methods and Techniques The course is conducted with practical laboratory studies along with theoretical presentations. Students develop models and design experiments on real data sets using deep learning libraries such as PyTorch or TensorFlow. The application of each topic progresses in the form of coding the relevant algorithm, performance analysis and comparison with the literature. With the project-based learning approach, students aim to solve a current problem in the field of computer vision by developing their own models.
Prerequisites and co-requisities None
Course Coordinator Associate Prof.Dr. Rifat Kurban - rifat.kurban@agu.edu.tr
Name of Lecturers Associate Prof.Dr. Rifat Kurban - rifat.kurban@agu.edu.tr
Assistants None
Work Placement(s) No

Recommended or Required Reading
Resources Computer Vision: Algorithms and Applications, 2nd ed.
Course Notes Computer Vision: Algorithms and Applications, 2nd ed.
2022 Richard Szeliski, The University of Washington
Documents Computer Vision: Algorithms and Applications, 2nd ed.
Assignments -
Exams -

Course Category
Mathematics and Basic Sciences %30
Engineering %40
Engineering Design %20
Social Sciences %0
Education %0
Science %0
Health %0
Field %10

Planned Learning Activities and Teaching Methods
Activities are given in detail in the section of "Assessment Methods and Criteria" and "Workload Calculation"

Assessment Methods and Criteria
In-Term Studies Quantity Percentage
Ödev 5 % 50
Sunum/Seminer 1 % 50
Total
6
% 100

 
ECTS Allocated Based on Student Workload
Activities Quantity Duration Total Work Load
Araştırma Ödevi 1 133 133
Ev Ödevi 5 10 50
Senkron Ders 14 3 42
Total Work Load   Number of ECTS Credits 7,5 225

 
Course Learning Outcomes: Upon the successful completion of this course, students will be able to:
NoLearning Outcomes
1 Analyze the architecture of deep learning-based computer vision models (e.g., CNNs, Transformers) and explain their structural components.
2 Select appropriate deep learning models for different visual tasks (e.g., object detection, segmentation, pose estimation) and design training strategies.
3 Evaluate the performance of deep learning-based vision solutions in terms of accuracy, complexity, and efficiency
4 Compare deep learning approaches to contemporary computer vision problems and develop original solution proposals.

 
Weekly Detailed Course Contents
WeekTopicsStudy MaterialsMaterials
1 Introduction to Deep Learning and Fundamental Concepts - -
2 Artificial Neural Networks and Feedforward Networks - -
3 Introduction to Convolutional Neural Networks (CNNs) - -
4 CNN Architectures and Visual Feature Learning - -
5 Image Classification and Transfer Learning - -
6 Object Detection: R-CNN, Fast R-CNN, Faster R-CNN - -
7 Object Detection: YOLO and SSD - -
8 Image Segmentation: FCN, U-Net - -
9 Semantic and Instance Segmentation - -
10 Pose Estimation and Keypoint Detection - -
11 Image Generation with Generative Adversarial Networks (GANs) - -
12 Visual Attention and Vision Transformer Models - -
13 Applied Project Development and Dataset Management - -
14 Project Presentations and Current Research Discussions - -
15 Final - -

 
Contribution of Learning Outcomes to Programme Outcomes
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11
All 4 5 4 4 4 3 4 1 2 2 3
C1 4 5 3 4 3 2 3 1 2 1 2
C2 3 5 4 4 4 4 4 1 2 2 2
C3 4 5 4 4 4 3 4 1 2 2 3
C4 5 5 4 5 5 3 4 2 3 2 3

  Contribution: 1: Very Slight 2:Slight 3:Moderate 4:Significant 5:Very Significant

  
  https://sis.agu.edu.tr/oibs/bologna/progCourseDetails.aspx?curCourse=77872&lang=en
Return to Classes