|
Departments >> Faculty of Engineering >> Department of Computer Science >>
|
Chair of Computer Science 5 (Pattern Recognition)
|
Algorithms and Data Structures for Medical Engineers [AuD-MT] -
- Lecturer:
- Peter Wilke
- Details:
- Vorlesung, 4 cred.h, ECTS: 5, für Anfänger geeignet, Frühstudium
- Dates:
- Mon, 16:15 - 17:45, Zoom-Webinar
Fri, 10:15 - 11:45, Zoom-Webinar
Beginn Vorlesung: 2. Nov. 2020, Beginn Übungen ab 9. Nov. 2020
- Fields of study:
- PF MT-BA 1
WF MT-MA 1-2
PF WINF-BA 1
PF DS-BA 1
- Prerequisites / Organisational information:
- Die Vorlesung wird als Aufzeichnung angeboten. Die Videos werden über das StudOn-Portal verteilt:
Kurs: AuD-MT WS 20/21
Ordner: Vorlesung VideosZu den Vorlesungszeiten wird eine Fragestunde angeboten:
Zoom-Webinar IDs
Montags: 976 9059 7167
Freitags: 934 0356 3371
Für die Teilnahme ist kein Passwort, aber login von FAU-Account aus notwendig.
Bitte den Zoom-Client verwenden, Browser-PlugIns werden nicht vollständig unterstützt. Das Passwort für den StudOn-Zugang wird, zusammen mit vielen weitern Informationen, in der 1. Vorlesung bekannt gegeben. AuD-MT besteht aus zwei Modulen:
AuD-MT-V, die Vorlesung, und
AuD-MT-UE, den Rechner- und Tafelübungen zur Vorlesung.
- Contents:
- Die Vorlesung AuD-MT richtet sich an Studierende der Studiengänge Medizintechnik, Wirtschafts-Informatik und weiteren. Es zählt dort zu den Grundlagenvorlesungen im Bereich Informatik.
Neben einer Einführung in die (objektorientierte) Programmierung in Java werden verschiedene Datenstrukturen wie verkettete Listen, Bäume und Graphen behandelt.
Ein weiterer Schwerpunkt liegt auf dem Entwurf von Algorithmen.
Dazu zählen Rekursion, Such- und Sortierverfahren, Graphalgorithmen, sowie die Aufwandsabschätzung von Algorithmen.
- Recommended literature:
- In der Vorlesung werden zu den einzelnen Kapiteln passende Lehrbücher vorgeschlagen.
- Keywords:
- Algorithmen, Datenstrukturen, Java
|
|
Datenbank Praxis [DBPraxis] -
- Lecturer:
- Peter Wilke
- Details:
- Vorlesung, 4 cred.h, ECTS: 5, Online-Kurs im StudOn
- Dates:
- Online-Kurs im StudOn
- Fields of study:
- WF INF-BA ab 4
WF INF-MA ab 1
WF INF-BA-V-SWE ab 4
WF INF-BA-V-DB ab 4
- Prerequisites / Organisational information:
- * Dies ist ein Online-Kurs , betreutes Eigenstudium! *
Keine formalen Voraussetzungen, grundlegende Kenntnisse im Bereich Datenbanken (zum Beispiel durch Besuch der Grundlagenvorlesungen KonzMod und IDB im Bachelor) werden empfohlen.
Der Kurs wird als Online-Kurs im Selbststudium angeboten. Die Kommunikation erfolgt per E-Mail und dem Forum im StudOn-Kurs. Ggf. wird ein Besprechungstermin vereinbart.
Der Kurs setzt die sichere Beherrschung einer Programmiersprache (z.B. Java) voraus, ebenso Erfahrung mit IDEs (Eclipse o. ä.). Erste Erfahrung im Umgang mit Mainframes (z/OS, TSO, ISPF) wäre hervorragend; z.B. Mainframe Programmierung I oder II.
- Contents:
- Datenbanken werden in fast jedem Unternehmen zur persistenten Datenspeicherung eingesetzt. Nach den Grundlagenvorlesungen im Bachelor, die die theoretische Einführung in die Datenbankwelt gegeben haben und die Basis für diesen Kurs bilden, wird in diesem Online-Kurs die praktische Erfahrung in der Arbeit mit einem Datenbanksystem in den Fokus gerückt. Der Online-Kurs ist so aufgebaut, dass es keine Vorlesungstermine und -videos gibt. Stattdessen kann der gesamte Inhalt in textueller Form über StudOn erarbeitet werden. Dies ermöglicht eine individuelle zeitliche Einteilung des Lernstoffs während der Vorlesungszeit.
Das in diesem Kurs verwendete Db2 for z/OS von IBM wird häufig im Enterprise-Umfeld eingesetzt. Insbesondere bei Banken, Versicherungsunternehmen und Softwarehäusern findet dieses Datenbanksystem Verwendung. Neben Oracle ist hier Db2 eines der weltweit am häufigsten eingesetzten Datenbanksysteme.
Die Kursinhalte umfassen:
Wiederholung der grundlegenden Konzepte aus den Bachelor-Pflichtvorlesungen
Einführung und Überblick über Db2 for z/OS
Administration von Db2 for z/OS
Programmzugriff auf Db2 for z/OS
Tools für Db2 for z/OS
Angewandte Aufgaben anhand eines Praxisbeispiels
- Recommended literature:
- Ist im StudOn-Kurs verlinkt
- Keywords:
- Mainframe, Programmierung, Programming, Administration, IBM, Datenbank, DB, Db2, Java, z, zOS
|
|
Deep Learning [DL] -
- Lecturer:
- Andreas Maier
- Details:
- Vorlesung, 2 cred.h, ECTS: 2,5, nur Fachstudium, Information regarding the online teaching will be added to the studon course
- Dates:
- Tue, 12:15 - 13:45, H4
- Fields of study:
- WPF INF-MA ab 1
WPF MT-MA-BDV 1
- Prerequisites / Organisational information:
- The following lectures are recommended:
Application via https://www.studon.fau.de/crs3317670.html
- Contents:
- Deep Learning (DL) has attracted much interest in a wide range of applications such as image recognition, speech recognition and artificial intelligence, both from academia and industry.
This lecture introduces the core elements of neural networks and deep learning, it comprises:
(multilayer) perceptron, backpropagation, fully connected neural networks
loss functions and optimization strategies
convolutional neural networks (CNNs)
activation functions
regularization strategies
common practices for training and evaluating neural networks
visualization of networks and results
common architectures, such as LeNet, Alexnet, VGG, GoogleNet
recurrent neural networks (RNN, TBPTT, LSTM, GRU)
deep reinforcement learning
unsupervised learning (autoencoder, RBM, DBM, VAE)
generative adversarial networks (GANs)
weakly supervised learning
applications of deep learning (segmentation, object detection, speech recognition, ...)
The accompanying exercises will provide a deeper understanding of the workings and architecture of neural networks.
- Recommended literature:
- Ian Goodfellow, Yoshua Bengio, Aaron Courville: Deep Learning. MIT Press, 2016
Christopher Bishop: Pattern Recognition and Machine Learning, Springer Verlag, Heidelberg, 2006
Yann LeCun, Yoshua Bengio, Geoffrey Hinton: Deep learning. Nature 521, 436–444 (28 May 2015)
- Keywords:
- deep learning; machine learning
|
|
Deep Learning Exercises [DL E] -
- Lecturers:
- Katharina Breininger, Florian Thamm, Felix Denzinger
- Details:
- Übung, 2 cred.h, ECTS: 2,5, nur Fachstudium, This course will be held online until the coronavirus pandemic is contained to such an extent that the Bavarian state government can allow face-to-face teaching again. Information regarding the online teaching will be added to the studon course
- Fields of study:
- WPF INF-MA ab 1
- Keywords:
- deep learning; machine learning
| | | Mon | 12:00 - 14:00 | 0.01-142 CIP | |
Breininger, K. Schröter, H. | |
| | Tue | 18:00 - 20:00 | 0.01-142 CIP | |
Breininger, K. Schröter, H. | |
| | Wed | 16:00 - 18:00 | 0.01-142 CIP | |
Breininger, K. Schröter, H. | |
| | Thu | 14:00 - 16:00 | 0.01-142 CIP | |
Breininger, K. Schröter, H. | |
| | Fri | 8:00 - 10:00 | 0.01-142 CIP | |
Breininger, K. Vesal, S. Schröter, H. | |
|
Introduction to Pattern Recognition [IntroPR] -
- Lecturers:
- Vincent Christlein, Mathias Seuret
- Details:
- Vorlesung, 3 cred.h, certificate, ECTS: 3,75, Information regarding the online teaching will be added to the studon course
- Dates:
- Wed, 14:15 - 15:45, H4
Fri, 12:15 - 13:45, H4
- Fields of study:
- WPF ME-BA-MG6 3-5
WPF MT-BA 5
WPF INF-BA-V-ME ab 5
WPF INF-BA-V-MI ab 5
WF CE-BA-TW ab 5
WPF INF-MA 1
WPF IuK-BA ab 5
WPF ME-MA-MG6 1-3
- Prerequisites / Organisational information:
- StudOn: https://www.studon.fau.de/crs2703226.html
- Keywords:
- Mustererkennung, Vorverarbeitung, Merkmalsextraktion, Klassifikation
|
|
Kolloquium Magnetic Resonance Imaging [MRI] -
- Lecturers:
- Andreas Maier, Armin Nagel, Frederik Laun, Moritz Zaiß, Sebastian Bickelhaupt, David Grodzki
- Details:
- Kolloquium, 2 cred.h, Das Kolloquium findet als Online-Veranstaltung statt. Zeit: Do 17:00 – 18:30 Uhr.
- Dates:
- Thu, 17:00 - 18:30, room tbd
- Fields of study:
- WPF INF-MA ab 1
- Keywords:
- magnetic resonance imaging, mri
|
|
Mainframe Programmierung [MainProg] -
- Lecturer:
- Peter Wilke
- Details:
- Vorlesung mit Übung, ECTS: 5
- Dates:
- Online-Kurs im StudOn, Zugang NUR über das VHB-Portal vhb.org !!!
- Fields of study:
- WF INF-MA ab 1
- Prerequisites / Organisational information:
- Der Kurs wird als Online-Kurs (StudOn) im Selbststudium angeboten. Die Anmeldung zum Kurs erfolgt über das StudOn-Portal, die Teilnehmerzahl ist begrenzt.
Die Kommunikation erfolgt per E-Mail und Forum.
Ggf. wird ein Besprechungstermin vereinbart. Der Kurs setzt die sichere Beherrschung einer Programmiersprache (z.B. Java, C/C++/C#, o.ä.) voraus, ebenso Erfahrung mit IDEs (z.B. Eclipse, VisualStudio, etc.)
- Contents:
- Der Begriff "Mainframe" bezeichnet grosse Rechenanlage, wie sie in der Wirtschaft für extrem grossen Anwendungen eingesetzt werden. Typische Branchen sind Banken und Versicherungen, aber auch Automobilhersteller und AI-Anwender.
Der Online-Kurs soll nun die Möglichkeit eröffnen, Erfahrungen mit der Programmierung eines Mainframes zu sammeln. Dazu gehören die elementaren Programmieraufgaben wie editieren, übersetzen, binden, laden, ausführen und debuggen, die anhand von Beispielen in der Programmiersprache CoBOL geübt werden.
Die Architektur der Mainframes werden sowohl aus Sicht der Rechnerarchitektur wie auch der Anwendersicht beleuchtet. Insbesondere werden die Virtualisierungsmöglichkeiten udn die gängigen Betriebssysteme wie z/OS und Linux auf den Mainframes behandelt.
Den Abschluss und Ausblick bildet die Datenhaltung und die Integration in die IT-Systemlandschaft.Inhalt:
0. Begrüßung und Einführung
1. CoBOL Programmierung
2. Einführung Mainframes
3. IBM Mainframe Architektur
4. z/OS
5. Anwendungsprogrammierung
6. Virtualisierung
7. Linux
8. Integration in die IT-Systemlandschaft
- Recommended literature:
- Wird über StudOn zur Verfügung gestellt.
- Keywords:
- Mainframe, Programmierung, Cobol, Fortran, z, zOS, CICS, REX, Rational
|
|
Mainframe@Home [MFH] -
- Lecturer:
- Peter Wilke
- Details:
- Vorlesung mit Übung, 4 cred.h, ECTS: 5, Online-Kurs im StudOn
- Dates:
- Online-Kurs im StudOn
- Fields of study:
- WPF INF-BA ab 4
WPF INF-MA ab 1
- Prerequisites / Organisational information:
- Der Kurs wird als Online-Kurs im StudOn im betreuten Selbststudium angeboten.
Zur Kommunikation mit den Kurs-Team steht ein Forum zur Verfügung. Außerdem kann Kontakt per E-Mail aufgenommen werden. Der Kurs setzt grundlegende Kenntnisse der Informatik voraus. Außerdem wird ein Verständnis für die Implementierung von Algorithmen benötigt. Anrechenbar für die Säulen:
softwareorientiert
anwendungsorientiert
- Contents:
- Großrechner sind das Herzstück der weltweiten IT-Landschaft. Durch die hohe Verfügbarkeit und geringe Ausfallquote werden Mainframes in sehr großen Firmen verwendet. Die Transaktionszahlen für die Datenverarbeitung sind bei diesen Unternehmen außerdem sehr hoch. Mit diesem Kurs soll Ihnen die Möglichkeit geboten werden, sich mit der Programmierung von Anwendungen für und der Arbeit mit Großrechner zu beschäftigen. Sie verwenden in diesem Kurs eine eigene Mainframe-Emulation auf Ihrem Rechner und arbeiten mit dieser in verschiedenen Übungsaufgaben.
Behandelt werden die folgenden Kapitel:
Einführung in das Thema Großrechner
Virtualisierung
Multiple Virtual Storage (MVS)
Common Business Oriented Language (Cobol)
Formula Translator (Fortran)
Restructured Extended Executor (Rexx)
Virtual Storage Access Method (VSAM)
Java und Unix auf dem Mainframe
- Recommended literature:
- Auf die Literatur wird in der jeweiligen Lerneinheit im StudOn hingewiesen.
- Keywords:
- Mainframe, Programmierung, zOS Betriebssystem, Rechner, Computer
|
|
Medical Image Processing for Diagnostic Applications (VHB course) [MIPDA] -
- Lecturers:
- Julian Hoßbach, Tristan Gottschalk, Lina Felsner, Stephan Seitz
- Details:
- Vorlesung, 4 cred.h, ECTS: 5
- Dates:
- to be determined
- Fields of study:
- WPF INF-MA ab 1
WPF INF-BA-V-ME ab 5
PF CE-MA-TA-IT ab 1
WPF IuK-MA-MMS-INF ab 1
WPF ICT-MA-MPS 1-4
WPF MT-MA-BDV ab 1
WPF MT-BA ab 5
WF CME-MA 1-4
- Prerequisites / Organisational information:
- Requirements: mathematics for engineering
Organization:
This is an online course of Virtuelle Hochschule Bayern (VHB).
Go to https://www.vhb.org to register to this course.
FAU students register for the written exam via meinCampus.
- Contents:
- Medical imaging helps physicians to take a view inside the human body and therefore allows better treatment and earlier diagnosis of serious diseases.
However, as straightforward as the idea itself is, so diversified are the technical difficulties to overcome when implementing a clinically useful imaging device. We begin this course by discussing all available modalities and the actual imaging goals which highly affect the imaging result. Some modalities produce very noisy results, but there are multiple other artifacts that show up in raw acquisition data and have to be dealt with. We address these issues in the chapter preprocessing and show how to compensate for image distortions, how to interpolate defect pixels, and finally correct bias fields in magnetic resonance images. The largest portion of this course covers the theory of medical image reconstruction. Here, from a set of projections from different viewing angles a 3-D image is merged that allows a definite localization of anatomical and pathological features. Following roughly the historical development of CT devices, we study the process from parallel beam to fan beam geometry and include a discussion of phantoms as a tool for calibration and image quality assessment. We then move forward and learn about reconstruction in 3-D. Since the system matrix often grows in dimensions such that many direct solvers become infeasible, we also discuss pros and cons of iterative methods. In the final chapter, image registration is introduced as the concept of computing the mapping that maps the content of one image to another. Two different acquisitions usually result in images that are at least rotated and translated against each other. Image registration forms the set of tools that we need to match certain image features in order to align both images for further processing, image improvement or image overlays.
- Keywords:
- Mustererkennung, Medizinische Bildverarbeitung
|
|
Pattern Recognition [PR] -
- Lecturer:
- Andreas Maier
- Details:
- Vorlesung, 3 cred.h, certificate, ECTS: 3,75, geeignet als Schlüsselqualifikation, This class will be given purely on fau.tv. Short videos will be posted on a regular schedule (not necessary the in-person time mentioned here at UnivIs)
- Dates:
- Thu, Fri, 10:15 - 11:45, H4
- Fields of study:
- WPF ME-BA-MG6 3-5
WPF MT-MA-BDV 1-3
PF IuK-MA-MMS-INF ab 1
PF ICT-MA-MPS 1-4
WPF CE-MA-INF ab 1
WPF INF-MA ab 1
WPF CME-MA ab 1
WF ASC-MA 1-4
WPF ME-MA-MG6 1-3
- Keywords:
- Mustererkennung, maschinelle Klassifikation
|
|
Pattern Recognition Exercises [PR E] -
- Lecturers:
- Stephan Seitz, Dalia Rodriguez Salas
- Details:
- Übung, 1 cred.h, ECTS: 1,25, nur Fachstudium, Information regarding the online teaching will be provided in the studon course.
- Fields of study:
- WPF ME-BA-MG6 3-5
WPF CE-MA-INF ab 1
WPF CME-MA ab 1
PF IuK-MA-MMS-INF ab 1
PF ICT-MA-MPS 1-4
WPF INF-MA ab 1
WPF MT-MA-BDV 1-3
WF ASC-MA 1-4
WPF ME-MA-MG6 1-3
- Keywords:
- Mustererkennung, Klassifikation
| | | Wed | 16:15 - 17:45 | 02.151-113 a CIP, 02.151-113 b CIP | |
Rodriguez Salas, D. Seitz, S. | |
| | Fri | 12:15 - 13:45 | Übung 3 / 01.252-128 | |
Rodriguez Salas, D. Seitz, S. | |
|
Projekt Computer Vision [ProjCV] -
- Lecturer:
- Vincent Christlein
- Details:
- Praktikum, ECTS: 10, This course will be held online until the coronavirus pandemic is contained to such an extent that the Bavarian state government can allow face-to-face teaching again
- Dates:
- Mon, 12:00 - 14:00, Übung 3 / 01.252-128, 00.156-113 CIP
- Fields of study:
- WPF INF-MA ab 1
WPF MT-MA ab 1
- Prerequisites / Organisational information:
- Basic knowledge of image processing is desirable. In the first session there will be a short recap on image representation and basic image filtering techniques. However, having visited lectures such as Introduction to Pattern Recognition (IntroPR) or Diagnostic Medical Image Processing (DMIP) might prove beneficial.
Please contact us if you have any questions. You can register via Studon (https://www.studon.fau.de/crs2949212.html) for the Computer Vision Project. During the semester lecture and exercise alternate on a weekly basis. Exercises are supervised and take place in one of the CIP pools. All exercises must be completed. You can get either 5 or 10 ECTS credits for this project. The following options are available:
5 ECTS (counts as: Hochschulpraktikum)
This option requires:
lectures (strongly recommended as they introduce the background required for the exercises)
exercises (in groups of 2 people) need to be finished on time
individual presentation about a state-of-the-art research paper at the end of the semester (graded if needed)
10 ECTS (counts as Hochschulpraktikum (5 ECTS) + Forschungspraktikum (5 ECTS), or Master Project Computer Science (10 ECTS))
lectures (strongly recommended as they introduce the background required for the exercises)
exercises (in groups of 2 people) need to be finished on time
individual coding/research project under supervision of a LME PhD student at the end of regular schedule (graded if needed)
Important: You cannot use the lecture/exercise part as a 5 ECTS research project (Forschungspraktikum). Please contact one of the PhD students at the lab if you need a research project.
- Contents:
- This project gives you the chance to learn about current computer vision topics and get practical experience in the field during the exercises.
Last semester, the following topics were covered:
- Keywords:
- Master Project, Pattern Recognition, Computer Vision
|
|
Seminar Deep Learning Theory & Applications [SemDL] -
- Lecturers:
- Vincent Christlein, Stefan Evert, Ronak Kosti
- Details:
- Seminar, 4 cred.h, graded certificate, ECTS: 5, nur Fachstudium, Information regarding the online teaching will be added to the studon course
- Dates:
- Wed, 10:15 - 11:45, 01.019
- Fields of study:
- WPF INF-MA 1
WPF MT-MA-BDV 1
WPF CE-MA-TA-MT 1
- Contents:
- Deep Neural Networks or so-called deep learning has attracted significant attention in the recent years.
They have had a transformative influence on Natural Language Processing (NLP) and Artificial Intelligence (AI), with numerous success stories recent claims of superhuman learning performance in certain tasks.
According to Young et al. (2017), more than 70% of the papers presented at recent NLP conferences made use of deep learning techniques.
Interestingly, the concept of Neural Networks inspired researchers already over generations since Minky's famous book (cf. http://en.wikipedia.org/wiki/Society_of_Mind ).
Yet again, this technology brings researchers to the believe that Neural Networks will eventually be able to learn everything (cf. http://www.ted.com/talks/jeremy_howard_the_wonderful_and_terrifying_implications_of_computers_that_can_learn ).
This year's main topic is: "Multi-Task Learning for Document Analysis",
i.e. we will analyze Documents of different nature (text, images, etc.)
by means of multi-task learning using different techniques, such as
Natural Language Processing, Handwriting recognition, etc.
- Recommended literature:
- Representation Learning: A Review and New Perspectives, Yoshua Bengio, Aaron Courville, Pascal Vincent, Arxiv, 2012.
Deep Learning, Yoshua Bengio, Ian Goodfellow, Aaron Courville, MIT Press
Goldberg, Yoav (2017). Neural Network Methods for Natural Language Processing. Number 37 in Synthesis Lectures on Human Language Technologies. Morgan & Claypool.
Young, Tom; Hazarika, Devamanyu; Poria, Soujanya; Cambria, Erik (2017). Recent trends in deep learning based natural language processing. CoRR, abs/1708.02709. http://arxiv.org/abs/1708.02709
Gradient-Based Learning Applied to Document Recognition, Yann Lecun, 1998
Dropout: A Simple Way to Prevent Neural Networks from Overfitting, Srivastava et al. 2014
Greedy layer-wise training of deep networks, Bengio, Yoshua, et al. Advances in neural information processing systems 19 (2007): 153.
Reducing the dimensionality of data with neural networks, Hinton et al. Science 313.5786 (2006): 504-507.
Training Deep and Recurrent Neural Networks with Hessian-Free Optimization, James Martens and Ilya Sutskever, Neural Networks: Tricks of the Trade, 2012.
Deep Boltzmann machines, Hinton et al.
Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion, Pascal Vincent et al.
A fast learning algorithm for deep belief nets, Hinton et al., 2006
ImageNet Classification with Deep Convolutional Neural Networks, Alex Krizhevsky, Ilya Sutskever, Geoffrey E Hinton, NIPS 2012
Regularization of Neural Networks using DropConnect, Wan et al., ICML
OverFeat: Integrated recognition, localization and detection using convolutional networks. Computing Research Repository, abs/1312.6229, 2013.
http://deeplearning.stanford.edu/wiki/index.php/UFLDL_Tutorial
http://deeplearning.net/tutorial/
Deep Learning Course on Coursera by Hinton
DL platform with GPU support: caffe, lasagne, torch etc.
Stanford University CS 224: Deep Learning for NLP (http://cs224d.stanford.edu )
University of Oxford: Deep Natural Language Processing (https://github.com/oxford-cs-deepnlp-2017/lectures )
- Keywords:
- deep learning; neural networks; machine learning; pattern recognition; natural language processing
|
|
Seminar Digital Pathology and Deep Learning [SemDP] -
- Lecturers:
- Katharina Breininger, Christian Marzahl, Andreas Maier, Samir Jabari, Ingmar Blümcke
- Details:
- Seminar, 2 cred.h, ECTS: 5, This course will be held online until the coronavirus pandemic is contained to such an extent that the Bavarian state government can allow face-to-face teaching again.
- Dates:
- Tue, 16:30 - 18:00, 02.133-113
- Fields of study:
- WPF MT-MA 1
WPF MT-MA-BDV 1
WPF INF-MA 1
- Contents:
- Pathology is the study of diseases and aims to deliver a fine-grained diagnosis to understand processes in the body as well as to enable targeted treatment. In this area, the opportunities for digital image processing are vast: While the need for precision medicine, i.e., taking into account various co-dependencies when formulating the best possible treatment for a patient, is high, the number of pathologists ist not increasing accordingly. Deep learning-based techniques can be used for different objectives in this scope. Examples include screening large microscopy images for specific rare events, providing visual augmentation with analysis data. Additionally, the availability of massive data collections, including genomics and further biological factors, can be utilized to determine specific information about diseases that were previously unavailable.
This seminar is offered to students of medicine as well as computer sciences and medical engineering and similar. Students will have to present a topic from this field in a short (30 min) and comprehensive presentation.List of topics:
Staining and special stains (including immunohistochemistry, enzyme-based dyes and tissue microarrays)
Current computational pathology
Knowledge/Feature fusion into a diagnosis
Histopathology quality control
Data sets as limiting factor - limits of current data sets
Large scale / clinical grade solutions
Computational and augmented tumor grading
In vivo microstructural analysis
Big data in pathology (multi-omics)
Histology image registration
Staining differences and stain normalization
Transfer learning and domain adaptation
Explainable AI
Virtual staining
Digital workflow in Germany vs. the world
Limits of digital pathology
|
|
Seminar Intraoperative Imaging and Machine Learning [IIML] -
- Lecturers:
- Holger Kunze, Katharina Breininger, Holger Keil
- Details:
- Seminar, 2 cred.h, ECTS: 5
- Dates:
- Wed, 8:30 - 10:00, 09.150
- Fields of study:
- WPF MT-MA-BDV ab 1
WPF INF-MA ab 1
WPF ICT-MA ab 1
WPF CE-MA-INF ab 1
- Contents:
- For many applications, techniques like deep learning allow for considerably faster algorithm development and allow to automate tasks that were performed manually in the past. In medical imaging, a large variety of time-consuming tasks that interfere with clinical workflows has the potential for automation. However, at the same time new challenges arise like data privacy regulations and ethics concerns.
In this seminar, we want to develop an application that allows for the automation of an X-ray based intraoperative planning or measurement procedure from a holistic perspective. To this end, we will invite a surgeon to explain the medical background and visit the operating room to understand the surgeons’ needs while performing the task. Having understood the underlying medical problem, we will look into topics of data privacy, code of ethics, prototype development, and UI design for surgeons. Furthermore, we will touch regulatory requirements necessary for releasing software to clinics.
At the end of the seminar, the students will have developed and documented a prototypical application for the indented intraoperative use case.
Students will be able to
visit an operation room, following the rules of such an environment
perform their own literature research on a given subject
independently research this subject according to data privacy and ethical standard
present and introduce the subject to their student peers
give a scientific talk in English according to international conference standards
describe their results in a scientific report
|
|
Seminar Meta Learning [SemMeL] -
- Lecturers:
- Andreas Maier, Patrick Krauß, Joachim Hornegger
- Details:
- Seminar, 2 cred.h, ECTS: 5, nur Fachstudium, This course will be held online until the coronavirus pandemic is contained to such an extent that the Bavarian state government can allow face-to-face teaching again. Information regarding the online teaching will be provided in the studon course.
- Dates:
- Mon, 8:15 - 9:45, KH 1.021
- Fields of study:
- WPF INF-MA ab 1
WPF MT-MA-BDV ab 1
WPF CE-MA-TA-MT ab 1
- Prerequisites / Organisational information:
- Registration via StudOn:
https://www.studon.fau.de/crs3330572.html
- Contents:
- Meta-learning refers to algorithms which aim to learn an aspect of a learning algorithm from data.
Examples of meta-learning methods include algorithms which design neural network architectures based on data, optimize the performance of a learning algorithm or exploit commonalities between tasks to enable learning from few samples on unseen tasks.
These methods hold the promise to automate machine learning even further than learning good representations from data by learning algorithms to learn even better representations.The seminar will cover the most important works which provide the cornerstone knowledge to understand cutting edge research in the field of meta-learning. Applications will include:
Learning from few samples
Automatically tuning neural network architectures
Determining appropriate equivariances
Disentangling causal mechanisms
- Recommended literature:
- Finn et al., "Model-agnostic meta-learning for fast adaptation of deep networks", ICML 2017
Zhou et al., "Meta-learning symmetries by reparameterization", Arxiv
Snell et al., "Prototypical networks for few-shot learning", Neurips 2017
Triantafillou et al., "Meta-dataset: A dataset of datasets for learning to learn from few examples", ICLR 2020
Vinyals et al., "Matching networks for one shot learning. ", Neurips 2016
Zoph et al. "Neural Architecture Search with Reinforcement Learning", Journal of Machine Learning Research 20 (2019)
Bengio et al., "A meta-transfer objective for learning to disentangle causal mechanisms", ICLR 2020
Santoro et al., "Meta-Learning with Memory-Augmented Neural Networks", ICML 2016
Ravi et al., "Optimization as a model for few-shot learning", ICLR 2016
Munkhdalai et al., "Meta Networks", ICML 2017
Sung et al. "Learning to Compare: Relation Network for Few-Shot Learning", CVPR 2018
Nichol et al. "On First-Order Meta-Learning Algorithms", Arxiv
- Keywords:
- algorithms; medical image processing
|
|
|
|