Biomedical Engineering Summer School

Biomedical Applications of Machine Learning

Gábor Kertész

The goal of the lectures is to provide an introduction to image recognition with neural networks, both is theory and in practice.

The course starts with a basic introduction to the fundamentals of image recognition using neural networks. On the practical side, Tensorflow and Keras is introduced as the most common frameworks used by machine learning engineers.

During the course the students will get an understanding on how Neural Networks are trained, evaluated, and what common problems could occur. After introducing Convolutional Neural Networks, classical architectures are analysed along with modern, state-of-the-art solutions and techniques, which improve the performance of neural networks.

Course topics:

• Regression and classification in fully connected networks
• Handling image data
• Convolutional Neural Networks
• Classic architectures and techniques
• State-of-the-art solutions and approaches to increase performance
• Hands-on with Tensorflow and Keras

Biomedical Image Processing

Dániel Kiss

This is an introductory course in the fundamentals of biological and medical imaging and image processing, including pre-processing and enhancement techniques, filtering, object detection, and image segmentation.

Participants will learn the basics of digital imaging and will gain experience in Python’s image processing abilities. We are also going to meet CellProfiler, an open-source application for batch processing, measuring, and analyzing images. We will see how to use them via real-life examples of various biological experiments covering the following topics.

Course topics:

• Fundamentals of image acquisition and image processing
• Intensity-based methods: histogram correction and thresholding
• Classical filters: smoothing, sharpening, edge enhancement
• Image segmentation and object identification
• Working with multi-channel images, measuring colocalization

Automated Machine Learning and Hyperparameter Optimization

Marcelo Vinícius Cysneiros Aragão

This course equips you with practical techniques to streamline your machine learning workflow. You'll gain a strong foundation in machine learning problems like classification, regression, and clustering.

The course dives into the world of automation with tools like scikit-learn, auto-sklearn, and Optuna. These tools can automate repetitive tasks in model development, potentially freeing up your time to focus on other aspects. You'll gain practical experience through hands-on exercises building and refining classification, regression, and clustering models using these tools.

By the end of the course, you'll be prepared to tackle real-world machine learning challenges. You'll develop the skills to potentially improve model development efficiency using AutoML and HPO. This could translate to time and resource savings.

Introduction to Project Management

Dr. Miklós Daróczi, Associate Professor

The main objective of the course is to deliver some theoretical and practical knowledge related mainly to the project planning, scheduling and budgeting. The module starts by defining the project and differentiating project management from general management. The project manager’s role, the project life cycle and the elements of project plan are also briefly discussed. The common formats of schedule the Gantt-charts and PERT/CPM networks, some methods of budgeting and cost estimating are also covered. Based on these techniques students will be able to participate in project planning and implementation.

Student teams will prepare a report and give a presentation on a special project. The purpose of this assignment is to deepen your knowledge on planning projects and to share that knowledge with the rest of the group. The topic of the project will be selected by the students.

Basics and Biomedical Applications of Microfluidics

Prof. Dr. Michael Schlüter; Dr. Diana Pinho

This Microfluidics course will give a basic overview of micro technologies and the corresponding manufacturing methods (e.g. photolithography, xurography, 3D printing). It furthermore addresses the study of basic and applied aspects of (bio)microfluidics, blood flow in microdevices, and biomedical applications (e.g. PCR-Chip, single cell analysis, plasma separation). This course will introduce participants to microfabrication technologies available for building microfluidic devices, with experimental practice. In the laboratory work, different microfluidic structures for (bio-)medical applications and blood flow studies will be developed, manufactured, and characterized.
 
Course topics:

•    Fundamentals of MEMS processing (frontend and backend processes, photolithographic process steps, thin film technology, etching technologies);
•    Fundamentals of Lab-on-a-Chip systems (LoC);
•    Advanced manufacturing methods e.g. in polymer materials and non-lithography methods;
•    Practice in manufacturing and characterization of a (micro-)fluidic system;
•    Development of a LoC (e.g. Assaychip, PCR-Chip…)
•    Fundamentals of biomicrofluidics and blood flow;
•    Microfluidic system applications (e.g. cell sorting, single cell analysis, plasma separation, organ-on-chip);

3D Modelling and Simulation of Biomechanical Structures

Alin Totorean, Eng, PhD, Sergiu Galatanu, Eng, PhD

The concept of personalized medicine has strongly evolved nowadays and is more and more used in clinical practice. For example, in orthopedics or cardiovascular medicine having patient specific models before interventions proves to be useful for physicians, since it allows them to plan the procedure in advance taking into consideration patient’s details. The process requires 3D models of the anatomical structures and further prototyping or performing computational simulations.

Amperometric Biosensors and Related Electronics

Dr. Cengiz Koçum, Prof. Dr. Dilek Çökeliler Serdaroğlu

Biosensors are used to detect biological or chemical markers (bio-chemicals, DNA, protein, cells etc.) in order for diagnosis, environmental monitoring and pharmaceutical research. This course represents the fundamental science and engineering concepts of designing biosensors through a common example. Biosensors have three main parts; biorecognition layer, transducers and electronics and they usually produce an electrical signal related to the specific analyte. In this course with lectures and practical exercises students will design and implement the main parts of an amperometric biosensor capable of glucose monitoring. Students will gain an insight of the complexities and design principles of a biosensor from scratch.

Course Outline:

• Introduction to biosensors and nanobiosensors
• Types of biorecognition elements
• Surface functionalizing and modification approaches for biosensors
• Transducer Basics: Sensors and Actuators
• Signal acquisition and conditioning principles
• Amperometric glucose biosensors
• Design and production of glucose sensing layer
• Signal conditioning for the amperometric glucose biosensor
• Electronic circuit and PCB design.

Instrumentation, Acquisition and Signal Processing for Biosignals

Dr. Mehmet Yüksekkaya

Biomedical signals are used to obtain crucial information about human health for diagnosis and treatment follow up. Acquiring biosignals from human body to a computer system is called biosignal acquisition and conditioning. For signal acquisition and conditioning, appropriate electronic instrumentation should be designed. In order to extract the crucial information from the acquired data, it should be processed.

This course is designed to be a practical introduction for data acquisition and processing. Students will acquire signals using adequate instrumentation. Further they apply data analysis. After this course, students are able to design a practical data acquisition system.

Course Synopsis: 

• Introduction to the design of analog circuits for signal conditioning for biosignal acquisition.
• Introduction to microcontroller-based system design.
• Practical design of a biosignal acquisition instrument.
• Introduction to processing of signals for feature extraction. 

Embedded Systems

Dr. Gerardo Marx Chávez-Campos

Nowadays, the “Embedded Systems” and its integration with the “Internet of Things” has become crucial for developing healthcare and industrial applications. Thus, to understand the developing process of apps on the embedded system that uses as the central operative system “Linux” are fundamental. Even more, this first approach to the most used operative system on servers, supercomputers, and single devices will guide the student to understand the complete integration of hardware electronics and code for developing from simple to complex tasks.

During the course, the student will realize by doing the fundamentals of how to work on the embedded board by accessing hardware, input/output digital ports, read the digital to analog converters. Then, the student will create an application to read and log data into structured frameworks. The registered data will be processed and analyzed using Python.

Course Synopsis: 

• Introduction to the beagle bone boards
• The beagle software
• Developing basic applications
• The Linux file system
• Interfacing electronics
• GPIO interfacing
• Internet of things
• A final project