Computer vision: image recognition and image analysis with artificial intelligence
.svg)
.svg)
.svg)
Machine Learning Engineer
1. basics of computer vision
- Why deep learning for computer vision tasks?
- Structure and function of neural networks
- Neural networks in image processing
2. introduction to TensorFlow and Keras
- Functions and cooperation of the frameworks
- Creating and training neural networks
3. convolutional neural networks (CNN)
- Architecture and functionality of CNN
- How is image data processed in CNN?
- Practical exercise: Performing image classification with CNN
4. image data augmentation in practice
- The concept of data augmentation
- Various augmentation techniques
- How do I improve my model with augmentation?
5. transfer learning in image classification
- Motivation and how transfer learning works
- Use of pre-trained models
- Fine-tuning pre-trained models for specific tasks
6. explainable image classification
- Motivation and challenges of Explainable AI
- Methods with which AI models can be explained
- Practical examples of application and implementation
7. introduction to PyTorch
- Overview of the PyTorch framework
- Comparison with TensorFlow and Keras
- A practical exercise with PyTorch
8. object recognition
- Challenges in object recognition
- The YOLO architecture and its application
- Practical exercise: Object recognition on your own data
9. image segmentation
- Overview of segmentation methods
- Image segmentation with CNN
- Image retrieval with embeddings
10. generative and multimodal AI
- Generative Adversarial Networks (GAN)
- Variational Autoencoder (VAE)
- Multimodal Large Language Models
- Connection of image and text data
A quick introduction to computer vision: You will understand the most important architectures - from CNNs to YOLO and vision transformers - and select the optimum model for each task.
Higher model quality: Data augmentation, transfer learning and explainable AI techniques increase precision and confidence in your results.
Well-founded technology decisions: Best-practice comparisons between TensorFlow/Keras and PyTorch, combined with deployment options from Cloud to Edge, give you a clear roadmap for your own projects.
Robust production applications: You will learn to recognize domain shifts, reliably detect small objects and minimize false alarms - essential for industrial, medical and retail applications.
Support for learning transfer: cloud lab, source code, Jupyter notebooks and deploy blueprint ensure transfer to your everyday work.
This training training takes place in a group of a maximum of 12 participants via Zoom.
In a cloud-based laboratory environment, you work directly in the browser. No further software needs to be installed.
Interactive Jupyter notebooks serve as learning material and working environment. They contain source code, documentation and links, while a powerful GPU server takes over the training of current models.
The course is held in German, the course materials are mostly available in English due to the focus on programming.
There is plenty of time for questions - individual support from the trainers is guaranteed.
You can access further materials in your personal learning environment.
developers developers, ML engineers, data scientists, engineers and consultants consultants who want to automatically evaluate image data or digitize visual inspections. Basic knowledge of Python is required. In-depth data science knowledge is helpful, but not essential.
.svg)
