arXiv subjects

Christoph Reich

Publications and source records attributed to Christoph Reich.

25 records · Page 2Linked to original sources

arXiv2021

OSS-Net: Memory Efficient High Resolution Semantic Segmentation of 3D Medical Data

Convolutional neural networks (CNNs) are the current state-of-the-art meta-algorithm for volumetric segmentation of medical data, for example, to localize COVID-19 infected tissue on computer tomography scans or the detection of tumour volumes in magnetic resonance imaging. A key limitation of 3D CNNs on voxelised data is that the memory consumption grows cubically with the training data resolution. Occupancy networks (O-Nets) are an alternative for which the data is represented continuously in a function space and 3D shapes are learned as a continuous decision boundary. While O-Nets are significantly more memory efficient than 3D CNNs, they are limited to simple shapes, are relatively slow at inference, and have not yet been adapted for 3D semantic segmentation of medical data. Here, we propose Occupancy Networks for Semantic Segmentation (OSS-Nets) to accurately and memory-efficiently segment 3D medical data. We build upon the original O-Net with modifications for increased expressiveness leading to improved segmentation performance comparable to 3D CNNs, as well as modifications for faster inference. We leverage local observations to represent complex shapes and prior encoder predictions to expedite inference. We showcase OSS-Net's performance on 3D brain tumour and liver segmentation against a function space baseline (O-Net), a performance baseline (3D residual U-Net), and an efficiency baseline (2D residual U-Net). OSS-Net yields segmentation results similar to the performance baseline and superior to the function space and efficiency baselines. In terms of memory efficiency, OSS-Net consumes comparable amounts of memory as the function space baseline, somewhat more memory than the efficiency baseline and significantly less than the performance baseline. As such, OSS-Net enables memory-efficient and accurate 3D semantic segmentation that can scale to high resolutions.

eess.IV↗

arXiv2021

Multi-StyleGAN: Towards Image-Based Simulation of Time-Lapse Live-Cell Microscopy

Time-lapse fluorescent microscopy (TLFM) combined with predictive mathematical modelling is a powerful tool to study the inherently dynamic processes of life on the single-cell level. Such experiments are costly, complex and labour intensive. A complimentary approach and a step towards in silico experimentation, is to synthesise the imagery itself. Here, we propose Multi-StyleGAN as a descriptive approach to simulate time-lapse fluorescence microscopy imagery of living cells, based on a past experiment. This novel generative adversarial network synthesises a multi-domain sequence of consecutive timesteps. We showcase Multi-StyleGAN on imagery of multiple live yeast cells in microstructured environments and train on a dataset recorded in our laboratory. The simulation captures underlying biophysical factors and time dependencies, such as cell morphology, growth, physical interactions, as well as the intensity of a fluorescent reporter protein. An immediate application is to generate additional training and validation data for feature extraction algorithms or to aid and expedite development of advanced experimental techniques such as online monitoring or control of cells. Code and dataset is available at https://git.rwth-aachen.de/bcs/projects/tp/multi-stylegan.

cs.CV↗

arXiv2021

Generation and Simulation of Yeast Microscopy Imagery with Deep Learning

Time-lapse fluorescence microscopy (TLFM) is an important and powerful tool in synthetic biological research. Modeling TLFM experiments based on real data may enable researchers to repeat certain experiments with minor effort. This thesis is a study towards deep learning-based modeling of TLFM experiments on the image level. The modeling of TLFM experiments, by way of the example of trapped yeast cells, is split into two tasks. The first task is to generate synthetic image data based on real image data. To approach this problem, a novel generative adversarial network, for conditionalized and unconditionalized image generation, is proposed. The second task is the simulation of brightfield microscopy images over multiple discrete time-steps. To tackle this simulation task an advanced future frame prediction model is introduced. The proposed models are trained and tested on a novel dataset that is presented in this thesis. The obtained results showed that the modeling of TLFM experiments, with deep learning, is a proper approach, but requires future research to effectively model real-world experiments.

eess.IV↗

arXiv2020

Attention-Based Transformers for Instance Segmentation of Cells in Microstructures

Detecting and segmenting object instances is a common task in biomedical applications. Examples range from detecting lesions on functional magnetic resonance images, to the detection of tumours in histopathological images and extracting quantitative single-cell information from microscopy imagery, where cell segmentation is a major bottleneck. Attention-based transformers are state-of-the-art in a range of deep learning fields. They have recently been proposed for segmentation tasks where they are beginning to outperforming other methods. We present a novel attention-based cell detection transformer (Cell-DETR) for direct end-to-end instance segmentation. While the segmentation performance is on par with a state-of-the-art instance segmentation method, Cell-DETR is simpler and faster. We showcase the method's contribution in a the typical use case of segmenting yeast in microstructured environments, commonly employed in systems or synthetic biology. For the specific use case, the proposed method surpasses the state-of-the-art tools for semantic segmentation and additionally predicts the individual object instances. The fast and accurate instance segmentation performance increases the experimental information yield for a posteriori data processing and makes online monitoring of experiments and closed-loop optimal experimental design feasible.

cs.CV↗

arXiv2020

Multiclass Yeast Segmentation in Microstructured Environments with Deep Learning

Cell segmentation is a major bottleneck in extracting quantitative single-cell information from microscopy data. The challenge is exasperated in the setting of microstructured environments. While deep learning approaches have proven useful for general cell segmentation tasks, existing segmentation tools for the yeast-microstructure setting rely on traditional machine learning approaches. Here we present convolutional neural networks trained for multiclass segmenting of individual yeast cells and discerning these from cell-similar microstructures. We give an overview of the datasets recorded for training, validating and testing the networks, as well as a typical use-case. We showcase the method's contribution to segmenting yeast in microstructured environments with a typical synthetic biology application in mind. The models achieve robust segmentation results, outperforming the previous state-of-the-art in both accuracy and speed. The combination of fast and accurate segmentation is not only beneficial for a posteriori data processing, it also makes online monitoring of thousands of trapped cells or closed-loop optimal experimental design feasible from an image processing perspective.

q-bio.QM↗

arXiv2020

Data Confidentiality In P2P Communication And Smart Contracts Of Blockchain In Industry 4.0

Increased collaborative production and dynamic selection of production partners within industry 4.0 manufacturing leads to ever-increasing automatic data exchange between companies. Automatic and unsupervised data exchange creates new attack vectors, which could be used by a malicious insider to leak secrets via an otherwise considered secure channel without anyone noticing. In this paper we reflect upon approaches to prevent the exposure of secret data via blockchain technology, while also providing auditable proof of data exchange. We show that previous blockchain based privacy protection approaches offer protection, but give the control of the data to (potentially not trustworthy) third parties, which also can be considered a privacy violation. The approach taken in this paper is not utilize centralized data storage for data. It realizes data confidentiality of P2P communication and data processing in smart contracts of blockchains.

cs.CR↗

arXiv2020

Agreements between Enterprises digitized by Smart Contracts in the Domain of Industry 4.0

The digital transformation of companies is expected to increase the digital interconnection between different companies to develop optimized, customized, hybrid business models. These cross-company business models require secure, reliable, and traceable logging and monitoring of contractually agreed information sharing between machine tools, operators, and service providers. This paper discusses how the major requirements for building hybrid business models can be tackled by the blockchain for building a chain of trust and smart contracts for digitized contracts. A machine maintenance use case is used to discuss the readiness of smart contracts for the automation of workflows defined in contracts. Furthermore, it is shown that the number of failures is significantly improved by using these contracts and a blockchain.

cs.CR↗