Professor for Image Analysis and Learning Heidelberg Collaboratory for Image Processing (HCI) Interdisciplinary Center for Scientific Computing (IWR) and Department of Physics and Astronomy Heidelberg University Im Neuenheimer Feld 205, 04.312; D-69120 Heidelberg Tel: +49 6221 5414803; Lab Manager Barbara Werner: +49 6221 5414833 fred.hamprecht@iwr.uni-heidelberg.de |

I develop machine learning algorithms for image analysis. More specifically, I am interested in principled methods that ingest images or video, and output "structured" predictions, such as the partitioning of an image into its constituent parts, or the tracking of all objects in a video. Most of our applications are in the life sciences; for instance, we try and segment all cells in a brain or a plant, or track virus components before assembly.

My focus is on methods that have a sound mathematical background, such as combinatorial optimization or algebraic graph theory, while being widely applicable and useful in practice. I attach particular importance to methods that can be trained with few examples, or by weak supervision, to empower colleagues from other disciplines to analyze their own data, conveniently. To the same end, I have launched the ilastik program for interactive machine learning (paper) that is now headed by Anna Kreshuk.

I enjoy, and feel most privileged, to be able to work on things unknown, and to teach the next generation of scientists and engineers.

I am proud that several past PhD students continue to serve in research and education, including

- Anna Kreshuk, Group Leader at EMBL
- Bjoern Andres, Professor at TU Dresden
- Bjoern Menze, Professor at TU Munich
- Bernhard Renard, Professor at HPI, Potsdam Universty

*and*nice.

## News

06.2020 Lab members have made crucial contributions to analysis of the "Kinderstudie" that has led to reopening of schools in the state of Baden-Wuerttemberg: press release.

10.2019 After a decennial effort, ilastik has become a potent tool for (bio-) image analysis. Development, including a "deep" version, is now headed by Anna Kreshuk. A new Perspective in Nature Methods summarizes the status quo.

09.2019 We have held the top spots in the ISBI and CREMI connectomics segmentation challenges almost continuously for three years. Kudos to Constantin Pape, Steffen Wolf, Thorsten Beier, Nasim Rahaman & co!

08.2019 Back from a fantastic sabbatical at Uppsala University. Thanks to Carolina Wahlby for hosting me!

09.2018 Thomas Hehn and I have won a best paper award in GCPR 2018 for our work on "Differentiable Decision Trees". Update: An extended version is now published by IJCV.

06.2018 Ullrich Koethe has been promoted to Adjunct Professor. Congratulations!

## Publications

List of publications by me (Google Scholar) or the entire lab (hosted locally).Selected publications:

**Multicut brings automated neurite segmentation closer to human performance.**Beier, T, Pape, C, Rahaman, N, Prange, T, Berg, S, Bock, D, Cardona, A, Knott, G W, Plaza, S M, Scheffer, L K, Köthe, U, Kreshuk, A and Hamprecht, F A.*Nature Methods*(2017) [10.1038/nmeth.4151]**A Comparative Study of Modern Inference Techniques for Structured Discrete Energy Minimization Problems**J. H. Kappes, B. Andres, F. A. Hamprecht, C. Schnörr, S. Nowozin, D. Batra, S. Kim, B. X. Kausler, T. Kröger, J. Lellmann, N. Komodakis, B. Savchynskyy, C. Rother.*IJCV 2015*[10.1007/s11263-015-0809-x | Technical Report | BibTeX]**Graphical Model for Joint Segmentation and Tracking of Multiple Dividing Cell**M. Schiegg, P. Hanslovsky, C. Haubold, U. Köthe, L. Hufnagel, F. A. Hamprecht.*Bioinformatics 2015*[10.1093/bioinformatics/btu764 | Technical Report | URL | BibTeX]**Globally Optimal Closed-Surface Segmentation for Connectomics**B. Andres, T. Kröger, K. L. Briggmann, W. Denk, N. Norogod, G. Knott, U. Köthe, F. A. Hamprecht*ECCV 2012*[10.1007/978-3-642-33712-3_56 | Technical Report | BibTeX]**Optimal lattices for sampling**H. R. Künsch, E. Agrell, F. A. Hamprecht*IEEE Transactions on Information Theory, (2005)*[10.1109/TIT.2004.840864]**Development and assessment of new exchange-correlation functionals**F. A. Hamprecht, A. J. Cohen, D. J. Tozer, N. C. Handy*Journal of Chemical Physics, (1998)*[10.1063/1.477267]

## Current Teaching

Summer semester 2020: Computer Vision: Foundations## Some video lectures

### Machine Learning for Computer Vision (Winter Term 2017)

- 1. Introduction
- 2. Undirected Probabilistic Graphical Models
- 2.1 MAP & Priors
- 2.2 Markov Random Fields
- 2.3 Gibbs Sampling
- 2.4 MRF as Integer Linear Program (I)
- 2.5 MRF as Integer Linear Program (II)
- 2.6 Tree-Shaped MRF
- 2.7 Belief Propagation
- 2.8 Gaussian MRF (I)
- 2.9 Gaussian MRF (II)
- 3. Neural Networks
- 3.1 Perceptrons
- 3.2 Back Propagation
- 3.3 Introduction to Deep Learning
- 3.4 Deep Learning Architectures
- 3.5 Natural Gradient Optimization (I)
- 3.6 Natural Gradient Optimization (II)
- 3.7 Combining Graphical Models & Neural Networks (I)
- 3.8 Combining Graphical Models & Neural Networks (II)
- 4. Directed Probabilistic Graphical Models
- 4.1 Reinforcement Learning
- 4.2 Policy Gradient
- 6.3 Robotics

### Pattern Recognition Video Lectures (Summer Term 2012)

Full playlist here.- 1.1 Applications of Pattern Recognition
- 1.2 k-Nearest Neighbors Classification
- 1.3 Probability Theory
- 1.4 Statistical Decision Theory
- 2.1 Pearson Correlation
- 2.2 Alternative Correl. Measures
- 2.3 Gaussian Graphical Models
- 2.4 Discriminant Analysis
- 3.1 Regularized LDA/QDA
- 3.2 Principal Component Analysis (PCA)
- 3.3 Bilinear Decompositions
- 4.1 History of Neural Networks
- 4.2 Perceptrons
- 4.3 Multilayer Perceptrons
- 4.4 The Projection Trick
- 4.5 Radial Basis Function Networks
- 5.1 Loss Functions
- 5.2 Linear Soft-Margin SVM
- 5.3 Nonlinear SVM
- 6.1 Kernels
- 6.2 One Class SVM
- 6.3 Random Forest
- 6.4 Random Forest Feature Importance
- 7.1 Least-Squares Regression
- 7.2 Optimum Experimental Design
- 7.3 Case Study: Functional MRI
- 7.4 Case Study: CT
- 7.5 Regularized Regression
- 8.1 Gaussian Process Regression
- 8.2 GP Regression: Interpretation
- 8.3 Gaussian Stochastic Processes
- 8.4 Covariance Function
- 9.1 Kernel Density Estimation
- 9.2 Cluster Analysis
- 9.3 Expectation Maximization
- 9.4 Gaussian Mixture Models
- 10.1 Bayesian Networks
- 10.2 Variable Elimination
- 10.3 Message Passing
- 10.4 State Space Models
- 11.1 The Lagrangian Method
- 11.2 Constraint Qualifications
- 11.3 Linear Programming
- 11.4 The Simplex Algorithm
- 12.1 StructSVM
- 12.2 Cutting Planes

**1 Introduction**

**2 Correlation measures, Gaussian Models**

**3 Dimensionality Reduction**

**4 Neural Networks**

**5 Support Vector Machines**

**6 Kernels, Random Forest**

**7 Regression**

**8 Gaussian Processes**

**9 Unsupervised Learning**

**10 Directed Graphical Models**

**11 Optimization**

**12 Structured Learning**

### Image Analysis Video Lectures (Summer Term 2013 / 2015).

Full playlist here, including some added material (scroll to last entries).- Introduction
- Human early vision
- Image representations
- Block Matching
- Texture Synthesis
- Non-Local Means for Image Denoising
- BM3D for image denoising
- Unitary transformations
- The Fourier Transform
- The Discrete Fourier Transform (DFT)
- 2D-DFT: Application to Images
- Fourier Transform
- Time Frequency Decompositions
- The Wavelet Transform
- Watershed
- Maximally Stable Extremal Regions (MSER)
- Mathematical Morphology
- Minkowski functionals
- Markov Random Fields (MRFs)
- Gaussian Markov Random Fields (GMRF)
- Intrinsic GMRFs (IGMRF)
- Factor Graphs
- Fields of Experts
- Discrete-Valued MRFs
- MAP inference via Integer Linear Programming (ILP)
- Integer Linear Programs (continued)
- Pseudo Boolean Functions (PBFs)
- Quadratic PBFs with submodular terms
- Max-Flow / Min-Cut
- Graph Cuts
- Introduction
- Example model: Tracking by assignment
- Structured Support Vector Machine (structSVM)
- Structured Learning: Applications
- Light Fields
- Coded Aperture Imaging
- Compressive Sensing

**1 Introduction**

**2 Patches in Image Analysis**

**3 Fourier Transformation**

**4 Wavelets**

**5 Images as Topographic Maps**

**6 Gaussian Random Fields**

**7 Fields of Experts, Discrete MRFs**

**8 Binary pairwise MRFs and Graph Cut**

**9 Structured learning**

**10 Light Fields and Compressive Sensing**