Friedrich-Alexander-Universität Erlangen
Lehrstuhl für Mustererkennung
Martensstraße 3
91058 Erlangen
This is software that our group developed and that we provide free of charge for academic purposes or under a free software license. We offer 
free datasets, too.
AQuAFo (GUI for Automated Quality Assessment of Retinal Fundus Photos) is a Graphical User Interface for Windows to estimate the quality of retinal fundus photos automatically.
More information is available here.
The Android PlotView library can be used to plot signals on Android devices. It is implemented as an Android control and has no external dependencies.
More information is available here.
This code accompanies the paper Camera Calibration: Active versus Passive Targets by Christoph Schmalz, Frank Forster and Elli Angelopoulou, (Optical Engineering, November 2011)
Active targets can be digital displays or e-paper. They have several advantages. Digital displays are nearly ubiquitous, so dedicated calibration targets with a fixed checkerboard or dot array pattern become unnecessary. Handling is easy as most displays have a tilt/swivel base to set up different target poses. Also, the target does not have to be in focus to obtain excellent calibration results, which is especially useful for calibrating cameras with a large field of view, since the target can be moved closer to the lens. Most importantly, active targets enable a higher localization accuracy for the calibration marks than traditional passive targets. This translates into a higher accuracy in applications like for example 3D reconstruction tasks.
See the paper and the included readme for further details.
Download link: 
ActiveCalib_v1.zip (C++, VC10 project included)
The main goal of 
CAVAREV is to enable an easy and objective comparison of different dynamic reconstruction algorithms.
The area of application is the 3-D and 4-D reconstruction of contrasted cardiac vessels, e.g. the coronary arteries using C-arm CT (rotational angiography). Various methods exist in literature with lots of nice results. However, the results can vary significantly between phantom and real clinical data.
Therefore, we provide:
For more information please visit www.cavarev.com.
This is a C++ library for general-purpose, unconstrained non-linear optimization on the GPU using NVIDA's CUDA. It implements the L-BFGS (“Limited-memory Broyden-Fletcher-Goldfarb-Shanno“) method, a popular Quasi-Newton optimization method with a low memory footprint.
More infrmation is available here
This is a C++ implementation of the framework proposed in our paper "GPU Accelerated Time-of-Flight Super-Resolution for Image-Guided Surgery". It employs maximum a posteriori (MAP) estimation to obtain an improved image of higher resolution from multiple low-resolution images. Our implementation accelerates super-resolution on the graphics processing unit (GPU) using NVIDIA's CUDA platform.
More information is available |   |
This MATLAB® tool supports comparison of edge sharpness in images, both interactively and in batch mode. More...
Epipolar Consistency has been shown to be a powerful tool in calibration and motion correction for flat panel detector CT (FD-CT) and other X-ray based applications. The epipolar consistency conditions (ECC) apply to projection data directly and can be used to correct for 3D parameters without 3D reconstruction. The computation of an epipolar consistency metric is real-time capable for small sets of images and can be parallelized for large sets of images. This is an open-source implementation in Octave/MATLAB, which can easily be ported to your favorite software environment.
Embedded microcontrollers are employed in an increasing number of applications as a target for the implementation of classification systems. This is true, for example, for the fields of sports, automotive, and medical engineering. However, important challenges arise when implementing classification systems on embedded microcontrollers, which is mainly due to limited hardware resources.
With the Embedded Classification Software Toolbox, we present a solution to the two main challenges, namely obtaining a classification system with low computational complexity and, at the same time, high classification accuracy. For the first challenge, we propose complexity measures on the mathematical operation and parameter level, because the abstraction level of the commonly used Landau notation is too high in the context of embedded system implementation. For the second challenge, we present a software toolbox that trains different classification systems, compares their classification accuracy, and finally analyzes the complexity of the trained system.
DynaSL is a collection of tools to analyze sampling strategies and reconstruction algorithms for T1-weighted DCE-MRI. It can be used to:
Further information and downloads can be found here.
FindTBB.cmake is a CMake module to automatically find the installation path of Intel's Threading Building Blocks.
Find more info about FindTBB.cmake at Google Code.
Gerbil is an open-source project intended for researchers working with multispectral or hyperspectral images, or researching and educating in color and reflectance. The software consists of a new interactive visualization tool targeted at multispectral and hyperspectral image data, and a toolbox of common algorithms, e.g. for segmentation.
For more information and download visit 
gerbilvis.org.
Hearty is an application for Android-based mobile devices that allows real-time electrocardiogram (ECG) monitoring and automated arrhythmia detection by analyzing ECG parameters.
More information is available here.
The Image Denoising Algorithms Archive contains a collection of commonly used and state-of-the-art denoising algorithms, implemented in Matlab. The algorithms were used and testet at the Pattern Recognition Lab.
Furthermore, data sets for the evaluation of denoising algorithms, which we used in our papers, may be downloaded.
Arrhythmia detection algorithms require the exact and instantaneous detection of fiducial points in the ECG signal. These fiducial points (QRS-complex, P- and T-waves) correspond to distinct cardiac contraction phases.
We implemented in Matlab three algorithms: one is based on a line fitting approach, an expeditious version of a Wavelet-based ECG delineation algorithm, and a fast naive fiducial points detection algorithm.
Download link: FiducialPointsDetectionAlgorithms.zip
Code snippets and code from our image forensics research. We put our generalization of the shift vector verification for copy-move forgery detection online, see the project page for details, proceed to the download of the code. Consider also to download the dataset.
The Java Speech Toolkit (JSTK) is an open source framework for speech processing, licensed under the GPL v3.0 and hosted at Google Code: http://code.google.com/p/jstk. It provides programs and API functionality for basic signal processing (IO, filtering), feature extraction (MFCC and others), Gaussian mixture model training (ML and MMIE), adaptation, small vocabulary size speech recognition and human rater evaluation.
Furthermore, JSTK provides visual components for the rapid development of speech annotation tools, such as JSTKTranscriber (package de.fau.cs.jstk.app.jstktranscriber) and Blitzscribe2 (package de.fau.cs.jstk.app.blitzscribe).
Java parallel optimization package (JPOP) is an open-source software library which allows the minimization of non-linear functions. The package provides interfaces to implement functions, gradients, and Hessians of arbitrary functions which can be minimized in a parallel manner. More info.
In this MATLAB toolbox, we provide the implementations of several state-of-the-art algorithms as well as novel methods developed in our projects on muti-frame super-resolution. The algorithms cover general-purpose reconstruction algorithms that exploit subpixel motion to gain super-resolved data and tailor-made solutions for specific applications (with focus on applications in medical imaging).
More information is available here
OCTSEG (Optical Coherence Tomography Segmentation and Evaluation GUI) is a
graphical user interface written in MATLAB for research purposes.
The retinal layers and the blood vessels of retinal OCT scans can be segmented. In addition, tools for the manual correction of the automated segmentations and for the visualization of the data are provided.
Further information is available here.
RabbitCT is an open-source benchmark for C-arm CT FBP reconstruction.
It features a software framework for measuring execution time and calculating reconstruction errors, and a clincal C-arm CT dataset with 496 projections.
For more information please visit rabbitct.com.
The Range Imaging Toolkit (RITK) provides a powerful yet intuitive software platform that facilitates the development of range image stream applications. RITK puts emphasis on real-time processing of range image streams and proposes the use of a dedicated pipeline mechanism. Furthermore, we introduce a powerful and convenient interface for range image processing on the graphics processing unit (GPU). Being designed thoroughly and in a generic manner, the toolkit is able to cope with the broad diversity of data streams provided by available RI devices and can easily be extended by custom range imaging sensors or processing modules. RITK is an open source project. It is an object-oriented, cross-platform toolkit written in C++, that can be used as a regular library as well as an application framework for rapid development.
More information:UnivIS2Typo3 project page at the LMEUnivIS2Typo3 repository at GoogleCode
ORCA-SPOT is a deep learning toolkit implemented in PyTorch and designed to segment interesting animal sounds (killer whale signals) out of large, noise-heavy, and heterogenous bioacoustic audio corpora. ORCA-SPOT can be adapted to other animal species. The code will be publicly available upon request only at the end of October 2019. For questions please contact Christian Bergler (christian.bergler@fau.de).
