Caenorhabditis elegans as an in vivo model organism provides the potential for higher throughput substance testing, leading to reduced animal testing, substance use, and experiment costs. In this work, white light and fluorescence images of closeup captures of C. elegans worms were used as a modality of measurement for the protein expression. To measure worm morphology and the effect of substances on the nematode’s behavior, fitness, and survivability relevant features will be extracted automatically. With automated segmentation and localization of worms in both modalities, important features can be extracted allowing conclusions on substance effects. For the segmentation, we used a Mask R-CNN to extract single worm instances and to allow the separation of close instances. Different effects on the training process and the combination of both image modalities were investigated. This results in a low MAPE and a high R^2 on unseen C. elegans images for important morphological and protein expression features such as mean intensity ( R^2 = 0.995), length ( R^2 = 0.952) and area ( R^2 = 0.983).

In histological preparations of tumors, an important criterion for evaluation of malignancy is the variation of nuclear size (anisokaryosis), which is traditionally estimated by pathologists. To improve reproducibility, nuclear size measurements (morphometry) can be performed. In this study [1] we developed a segmentation-based nuclear morphometry algorithm and compared its prognostic value with pathologist’s estimates (routine method) and manual morphometry (gold standard method).

Microcomputed tomography (µCT) is an essential tool for analyzing trabecular bone microarchitecture, yet its resolution is constrained by object size and acquisition time. To overcome these limitations, we implement a deep-learning-based super-resolution (SR) approach that enhances µCT image resolution while significantly reducing scan durations. Dry isolated tarsal bones (intermediate cuneiform) from 20 specimens were scanned using µCT at two resolutions, 80 µm voxel size (low resolution, LowRes) and 20 µm voxel size (high resolution, HiRes). Aligned LowRes and HiRes µCT data served as training data for SR reconstruction. In this study, we compare five SR models: 2D U-Net+ +, 3D SRCNN, 3D FSRCNN, 3D U-Net and a modified 3D U-Net model trained with a combined learned perceptual image patch similarity (LPIPS) and structural similarity (SSIM) loss function. The focus of this contribution is the application of these models based on real µCT data, rather than synthetically degraded images. Models were trained to learn volumetric representations for accurate restoration of trabecular bone microstructure. To assess SR image quality, we computed three image quality metrics (peak signal-to-noise ratio, SSIM and LPIPS) and evaluated bone morphometric parameters, i.e. average trabecular thickness (Tb.Th.) and bone volume fraction (BV/TV), across 95 regions of interest (ROI). RMSE was calculated for LowRes data and each SR model relative to HiRes data to quantify prediction accuracy. The results demonstrate that the 3D U-Net (LPIPS & SSIM) model achieves the highest reconstruction accuracy, yielding the lowest RMSE values (12.93 µm for Tb.Th. and 1.3 % for BV/TV), outperforming all other SR models in our evaluation. Compared to standard low-resolution µCT, our approach reduces scan time from 58 min to 7 min per sample while preserving trabecular morphology with high fidelity. These results demonstrate the effectiveness of perceptual loss-based SR to real µCT data for morphological analysis, ensuring accurate trabecular reconstruction and mitigating overestimation artifacts caused by LowRes imaging and partial volume effects. Integrating SR with real µCT scans offers a promising strategy to reduce scan time to improve throughput in bone imaging workflows. Future work will expand the training dataset to enhance model generalization across diverse bone structures and imaging conditions.

The integration of deep learning-based tools into diagnostic workflows is increasingly prevalent due to their efficiency and reproducibility in various settings. We investigated the utility of automated nuclear morphometry for assessing nuclear pleomorphism (NP), a criterion of malignancy in the current grading system in canine pulmonary carcinoma (cPC), and its prognostic implications. We developed a deep learning-based algorithm for evaluating NP (variation in size, i.e., anisokaryosis and/or shape) using a segmentation model. Its performance was evaluated on 46 cPC cases with comprehensive follow-up data regarding its accuracy in nuclear segmentation and its prognostic ability. Its assessment of NP was compared to manual morphometry and established prognostic tests (pathologists’ NP estimates (n = 11), mitotic count, histological grading, and TNM-stage). The standard deviation (SD) of the nuclear area, indicative of anisokaryosis, exhibited good discriminatory ability for tumor-specific survival, with an area under the curve (AUC) of 0.80 and a hazard ratio (HR) of 3.38. The algorithm achieved values comparable to manual morphometry. In contrast, the pathologists’ estimates of anisokaryosis resulted in HR values ranging from 0.86 to 34.8, with slight inter-observer reproducibility (k = 0.204). Other conventional tests had no significant prognostic value in our study cohort. Fully automated morphometry promises a time-efficient and reproducible assessment of NP with a high prognostic value. Further refinement of the algorithm, particularly to address undersegmentation, and application to a larger study population are required.

The recently developed high-throughput system for cell spheroid generation (SpheroWell) is a promising technology for cost- and time-efficient in vitro analysis of, for example, chondrogenic differentiation. It is a compartmental growth surface where spheroids develop from a cell monolayer by self-assembling and aggregation. In order to automatize the analysis of spheroids, we aimed to develop imaging software and improve the localization of cell compartments and fully formed spheroids. Our workflow provides automated detection and localization of spheroids in different formation stages within Petri dishes based on images created with a low-budget camera imaging setup. This automated detection enables a fast and inexpensive analysis workflow by processing a stack of images within a short period of time, which is essential for the extraction of early readout parameters. Our workflow combines image processing algorithms and deep learning-based image localization/segmentation methods like Mask R-CNN and Unet++. These methods are refined by an evolution strategy for automated grid detection, which is able to improve the overall segmentation and classification quality. Besides the already pre-trained neural networks and predefined image processing parameters, our evolution-based post-processing provides the required adaptability for our workflow to deliver a consistent and reproducible quality. This is especially important due to the use of a low-budget imaging setup with various light conditions. The to-be-detected objects of the three different stages show improved results using our evolutionary post-processing for monolayer and starting aggregation with Dice coefficients of 0.7301 and 0.8562, respectively, compared with the raw scores of 0.2879 and 0.8187. The Dice coefficient of the fully formed spheroids in both cases is 0.8829. With our algorithm, we provide automated analyses of cell spheroid by self-assembling in SpheroWell dishes, even if the images are created using a low-budget camera setup.

Variation in nuclear size and shape is an important criterion of malignancy for many tumor types; however, categorical estimates by pathologists have poor reproducibility. Measurements of nuclear characteristics can improve reproducibility, but current manual methods are time-consuming. The aim of this study was to explore the limitations of estimates and develop alternative morphometric solutions for canine cutaneous mast cell tumors (ccMCTs). We assessed the following nuclear evaluation methods for accuracy, reproducibility, and prognostic utility: (1) anisokaryosis estimates by 11 pathologists; (2) gold standard manual morphometry of at least 100 nuclei; (3) practicable manual morphometry with stratified sampling of 12 nuclei by 9 pathologists; and (4) automated morphometry using deep learning-based segmentation. The study included 96 ccMCTs with available outcome information. Inter-rater reproducibility of anisokaryosis estimates was low (k = 0.226), whereas it was good (intraclass correlation = 0.654) for practicable morphometry of the standard deviation (SD) of nuclear size. As compared with gold standard manual morphometry (area under the ROC curve [AUC] = 0.839, 95% confidence interval [CI] = 0.701-0.977), the prognostic value (tumor-specific survival) of SDs of nuclear area for practicable manual morphometry and automated morphometry were high with an AUC of 0.868 (95% CI = 0.737-0.991) and 0.943 (95% CI = 0.889-0.996), respectively. This study supports the use of manual morphometry with stratified sampling of 12 nuclei and algorithmic morphometry to overcome the poor reproducibility of estimates. Further studies are needed to validate our findings, determine inter-algorithmic reproducibility and algorithmic robustness, and explore tumor heterogeneity of nuclear features in entire tumor sections.

In the field of image analysis used for diagnostic processes, domain shifts constitute a significant obstacle. Domain shifts lead to an incompatibility of an otherwise well-performing AI model for image segmentation. Accordingly, if two different machines image the same tissue, the model may provide better results for one of the two images depending on the similarity of the image data compared to the training data for generating the AI model. In this paper, we analyzed how the input images of a neural network have to be adapted to provide better segmentation results for images which are previously not compatible with the used model. Therefore, we developed two approaches to increase a model’s segmentation quality for a dataset with initially poor results. The first approach is based on heuristic optimization and creates a set of image processing algorithms for the data adaptation. Our algorithm selects the best combination of algorithms and generates the most suitable parameters for them regarding the resulting segmentation quality. The second approach uses an additional neural network for learning the incompatible dataset’s recoloring based on the resulting segmentation quality. Both methods increase the segmentation quality significantly without the need for changes to the segmentation model itself.

Introduction Nuclear pleomorphism (variation in size and shape) is one of the prognostic parameters used to grade canine pulmonary carcinomas. Automated measurements of nuclear morphology may provide prognostic advantages over categorial estimations by pathologists.

The computation of assembly tolerance information is necessary to fulfill robust design requirements. This assembly is computationally costly, with current calculations taking several hours. We aim to identify surrogate models for predicting degrees of freedom within a tolerance chain based on point connections between assembly components. Thus, replacing part of the current computation workflow and consequently reduce computation time. We use manufacturing tolerances set by norms and industrial standards to identifly these surrogate models, which define all relevant features and resulting output variables. We use black-box modeling methods (artificial neural networks and gradient boosted trees), as well as white-box modeling (symbolic regression by genetic programming). We see that these three models can reliably predict the degrees of freedom of a tolerance chain with high accuracy (R2 > 0.99).

Micropatterning in living cells is used for the analysis of protein-protein interactions. The quantitative analysis of images produced within this process is time-consuming and non-trivial task. For the simplification and speedup of such analyses, we describe a method for fully automated analysis of micro-patterned cells in fluorescence microscopy images. An approach based on an evolution strategy allows the grid extraction of the assays to estimate the pattern on the cells. We outline a workflow for the segmentation of these patterned cells based on a Unet. We also show the efficiency of different data augmentations applied to different patterning setups. A Dice score of 0.89 with 3 µm patterns and 0.79 with 1 µm patterns could be achieved. As we demonstrate in this study, we can provide thorough micropatterning studies, by automating the cell segmentation process.

Introduction Anisokaryosis is a prognostic criterion for many tumor types. Although it is traditionally estimated by pathologists, measurements have advantages for statistical evaluation and may improve reproducibility. The aim of the study was to compare these two methods in canine cutaneous mast cell tumors (ccMCT).

Variation in nuclear size and shape is an important criterion of malignancy for many tumor types; however, categorical estimates by pathologists have poor reproducibility. Measurements of nuclear characteristics (morphometry) can improve reproducibility, but manual methods are time consuming. In this study, we evaluated fully automated morphometry using a deep learning-based algorithm in 96 canine cutaneous mast cell tumors with information on patient survival. Algorithmic morphometry was compared with karyomegaly estimates by 11 pathologists, manual nuclear morphometry of 12 cells by 9 pathologists, and the mitotic count as a benchmark. The prognostic value of automated morphometry was high with an area under the ROC curve regarding the tumor-specific survival of 0.943 (95% CI: 0.889 - 0.996) for the standard deviation (SD) of nuclear area, which was higher than manual morphometry of all pathologists combined (0.868, 95% CI: 0.737 - 0.991) and the mitotic count (0.885, 95% CI: 0.765 - 1.00). At the proposed thresholds, the hazard ratio for algorithmic morphometry (SD of nuclear area $\geq 9.0 \mu m^2$) was 18.3 (95% CI: 5.0 - 67.1), for manual morphometry (SD of nuclear area $\geq 10.9 \mu m^2$) 9.0 (95% CI: 6.0 - 13.4), for karyomegaly estimates 7.6 (95% CI: 5.7 - 10.1), and for the mitotic count 30.5 (95% CI: 7.8 - 118.0). Inter-rater reproducibility for karyomegaly estimates was fair ($\kappa$ = 0.226) with highly variable sensitivity/specificity values for the individual pathologists. Reproducibility for manual morphometry (SD of nuclear area) was good (ICC = 0.654). This study supports the use of algorithmic morphometry as a prognostic test to overcome the limitations of estimates and manual measurements.

Extracellular vesicles (EV) enable cell-to-cell communication in the body of an organism and carry significant potential in the medical field as disease indicators, tissue regeneration, and drug carriers. We developed two workflows to analyze EV data based on microscopy images reliably. The first workflow enables determining the total number of fluorophores per EV by analyzing the photobleaching step counts in a stepwise photobleaching experiment recorded through fluorescence microscopy. Furthermore, we present a workflow for the quality assessment of EV populations through multimodal imaging. Thus, enabling quantification of the purification quality (differentiation between EVs and other components) and the labeling ratio. Both workflows have shown excellent results on various data sets and under various conditions.

: Cell Spheroids are of high interest for clinical cell applications and cell screening. To allow the extraction of early readout parameters a high amount of image data of petri dishes is created. To support automated analyses of spheroids in petri dish images we present a method for analysing and quantification of spheroids in its development stages. The algorithm is based on multiple image processing algorithms and neural networks. With an evolutionary strategy, engraved grid cells on petri dish are extracted and on top a Unet is used for the segmentation and quantification of different cell compartment states. The measured f1-scores for the different states are 0.77 for monolayer grid cells, 0.86 for starting formation grid cells and 0.85 for spheroids. As we describe in this study we can provide thorough analyses of cell spheroid in petri dishes, by automating the quantification process.

Histopathological examination of tissue samples is essential for identifying tumor malignancy and the diagnosis of different types of tumor. In the case of lymphoma classification, nuclear size of the neoplastic lymphocytes is one of the key features to differentiate the different subtypes. Based on the combination of artificial intelligence and advanced image processing, we provide a workflow for the classification of lymphoma with regards to their nuclear size (small, intermediate, and large). As the baseline for our workflow testing, we use a Unet++ model trained on histological images of canine lymphoma with individually labeled nuclei. As an alternative to the Unet++, we also used a publicly available pre-trained and unmodified instance segmentation model called Stardist to demonstrate that our modular classification workflow can be combined with different types of segmentation models if they can provide proper nuclei segmentation. Subsequent to nuclear segmentation, we optimize algorithmic parameters for accurate classification of nuclear size using a newly derived reference size and final image classification based on a pathologists-derived ground truth. Our image classification module achieves a classification accuracy of up to 92% on canine lymphoma data. Compared to the accuracy ranging from 66.67 to 84% achieved using measurements provided by three individual pathologists, our algorithm provides a higher accuracy level and reproducible results. Our workflow also demonstrates a high transferability to feline lymphoma, as shown by its accuracy of up to 84.21%, even though our workflow was not optimized for feline lymphoma images. By determining the nuclear size distribution in tumor areas, our workflow can assist pathologists in subtyping lymphoma based on the nuclei size and potentially improve reproducibility. Our proposed approach is modular and comprehensible, thus allowing adaptation for specific tasks and increasing the users’ trust in computer-assisted image classification.

The quality of freshly harvested herbs is affected by several crucial factors, such as weather, tillage, fertilization, drying, and the harvesting process, e.g. Our main goal is to learn models that are able to predict spore contaminations in different types of herbs on the basis of information about the harvesting process, transport conditions, drying, and storage conditions. This shall enable us to identify optimal processing parameters, which will allow more effective and cost efficient contamination prevention. Using machine learning, we have generated ensembles of models that predict the risk for spore contamination on the basis of harvest processing parameters. The training information about contamination in herbs is given as results of laboratory analysis data. We applied different modeling algorithms (random forests, gradient boosting trees, genetic programming, and neural networks). In this paper we report on modeling results for yeast and mold contaminations in peppermint and nettle; e.g., for yeast contamination in peppermint we obtained models with 78.13% accuracy. Additionally, we use descriptive statistics to identify those parameters that have a statistically significant influence on the contamination; for example, our analysis shows that there seems to be a relationship between mold in peppermint and information about harrowing and the growth height (p = 0.001).

In this paper, we present an heterogeneous ensemble modeling approach to learn predictors for yeast contamination in freshly harvested peppermint batches. Our research is based on data about numerous parameters of the harvesting process, such as planting, tillage, fertilization, harvesting, drying, as well as information about microbial contamination. We use several different machine learning methods, namely random forests, gradient boosting trees, symbolic regression by genetic programming, and support vector machines to learn models that predict contamination on the basis of available harvesting parameters. Using those models we form model ensembles in order to improve the accuracy as well as to reduce the false negative rate, i.e., to oversee as few contaminations as possible. As we summarize in this paper, ensemble modeling indeed helps to increase the prediction accuracy for our application, especially when using only the best models. The final prediction accuracy as well as other statistical indicators such as false negative rate and false positive rate depend on the choice of the discrimination threshold; in the optimal case, model ensembles are able to predict yeast contamination with 65.91% accuracy and only 19.15% of the samples are false negative, i.e., overseen contaminations.

The aim of the article is to build up qualitative evidence that CBRN training, which requires quick and prompt reactions from first responders, should involve simulation. The authors rely on a case study, the European Sensor System for CBRN Applications EU-SENSE project, which has to provide an adaptable and multipurpose hazard detection solution for enhancing situational awareness. A dedicated training mode on how to use the device by CBRN practitioners assumes that life-threatening substances are present in a real-life operation. Therefore, in order to provide lifelike experiences in a controlled environment, a simulation-based activities can be proposed. The analysis of CBRN training requirements and sample hypothetical scenarios, which complement the case study, proves that CBRN mission-specific training requires the design of a range of simulationbased activities that help learners acquire safe behavioural competences and at the same time examine their performance. They will be also conducive to developing decision-making and communication skills and improving situational awareness. In a contaminated environment, these skills are necessary for survival. This study contributes to the growing body of research on CBRN training concerning newly developed devices, which enhance the work of CBRN practitioners.