Surgical gauze is an essential part of surgical procedures, which is primarily used for controlling bleeding and absorbing bodily fluids. The post-surgical retention of gauze can lead to serious complications in the patient's health and necessitate additional surgery for gauze removal. In the wake of data scarcity, the research on gauze segmentation on the real-world surgical data remains underexplored. In this work, we investigate the use of deep learning methods for gauze segmentation in robotassisted minimally invasive abdominal surgeries, utilizing an inhouse surgical dataset prepared at a university hospital. The training data reflects a realistic surgical setting and extensive diversity in spatial, morphological, and visual attributes of three different gauze categories. We have investigated prevalently used segmentation architectures, including CNN-based, transformer-based, and hybrid architectures, to provide a proof-of-concept for gauze segmentation in a realistic setting. Besides, we investigate the influence of additional sub-optimally annotated, auto-tracked segmentation masks to address the bottleneck of data scarcity and performance optimization. Our results demonstrate the efficacy of real-world data to counter the main challenge reported by prior works - the trade-off between blood presence and gauze detection. The incorporation of auto-track annotations enables performance enhancements, particularly in generic cases. The integration of effective segmentation approaches will benefit robotguided surgical procedures and various downstream applications by providing a precise delineation of foreign objects, enhancing patient safety and surgical outcomes.

Understanding the dynamics of blood flow is crucial in the context of cardiovascular health and disease. The dynamics of the blood flow can be a significant parameter for the development of decision support systems to enable early detection and accurate diagnosis of coronary artery diseases. Uncovering the underlying dynamics from high-dimensional data generated from experiments is a highly complex problem at the intersection of artificial intelligence and applied mathematics. Deep Hidden Physics Models can be used to learn the underlying dynamics without additional physical knowledge.In this work, the potential of Deep Hidden Physics Models to model the clinically relevant dynamics of blood flow is investigated. The experiments consider the use case of stenosis in two-dimensional spatial space. Based on the learned dynamics, the concentration field can be approximated accurately, indicating that the dynamics are learned correctly. Additionally, we examine the capability of the model to extrapolate the learned dynamics for unknown time intervals.

Deep Learning is proving its immense potential in medical image processing. However, noisy label data can weaken the generalization ability of the model and cause significant performance degradation. This issue can be more prevalent in multi-label classification tasks, since their annotation is more challenging. With too many labels, human annotators may have difficulties mentioning all possible classes, which leads to an increased number of false negative labels. Single Positive Multi-Label (SPML) training deals with the most severe version of this problem, in which for each sample only one positive label is available. All other labels are not observed, i.e. not confirmed as positive or negative. While SPML has already achieved good results for the multi-label detection of objects, its impact on the extremely pertinent medical imaging use case has not yet been explored. In this work, we therefore investigate the performance of state-of-the-art SPML loss functions in the analysis of chest X-rays, both on the public CheXpert and on an in-house data set of the University Hospital Bonn. In addition, we propose our new SPML loss functions, the Generalized Assume Negative and Implicit Weighting Assume Negative loss, which increase the mean average precision by up to 4.6% compared to the popular binary cross-entropy loss.

This paper investigates a data-and knowledge-driven approach to automatically analyze lumbar MRI scans.The dataset used is an in-house dataset of 142 sagital lumbar spine images from German radiology practices of the evidia GmbH.We implement state-of-the-art deep learning methods to segment the individual vertebral bodies.Overall, a very accurate segmentation performance of 97% Dice Score was achieved.Based on this segmentation, pathologically relevant distances are calculated using rule-based computer vision methods.We focus on the anterior, posterior and middle height of a vertebra and the anterior and posterior distances between two lumbar vertebrae.We demonstrate the clinical value of this approach through a quantitative and qualitative result analysis.

Radiologists commonly conduct chest X-rays for the diagnosis of pathologies or the evaluation of extrathoracic material positions in intensive care unit (ICU) patients.Automated assessments of radiographs have the potential to assist physicians by detecting pathologies that pose an emergency, leading to faster initiation of treatment and optimization of clinical workflows.The amount and quality of training data is a key aspect for developing deep learning models with reliable performance.This work investigates the effects of transfer learning on public data, automatically generated data labels and manual data annotation on the classification of ICU chest X-rays of the University Hospital Bonn.

Chest X-Ray imaging is one of the most common radiological tools for detection of various pathologies related to the chest area and lung function. In a clinical setting, automated assessment of chest radiographs has the potential of assisting physicians in their decision making process and optimize clinical workflows, for example by prioritizing emergency patients.Most work analyzing the potential of machine learning models to classify chest X-ray images focuses on vision methods processing and predicting pathologies for one image at a time. However, many patients undergo such a procedure multiple times during course of a treatment or during a single hospital stay. The patient history, that is previous images and especially the corresponding diagnosis contain useful information that can aid a classification system in its prediction.In this study, we analyze how information about diagnosis can improve CNN-based image classification models by constructing a novel dataset from the well studied CheXpert dataset of chest X-rays. We show that a model trained on additional patient history information outperforms a model trained without the information by a significant margin.We provide code to replicate the dataset creation and model training.