5805 research outputs found
Sort by
Machine learning model to classify chronic leg wounds and identify pyoderma gangrenosum
No full textStudy objectives
Chronic wounds represent a significant economic and personal burden. For their successful treatment, the causes must be known and treated. Wounds caused by pyoderma gangrenosum (PG), a rare inflammatory skin disease, are often misdiagnosed. This study, therefore, aims to develop a machine learning model capable of differentiating PG from other wound types, focusing on chronic leg wounds to address this diagnostic challenge.
Methods
We used 3674 wound photographs from three specialised wound centres with the four most common types of foot and leg ulcers and the rare inflammatory differential diagnosis PG. The convolutional neural network classifier ConvNeXt ‘B’ was pretrained on LAION2B, ImageNet12k and ImageNet 1k before being trained and fine-tuned on an 85:15 train, validation split.
Results
The model achieved an overall high accuracy in multiclass classification of the chronic wounds (unbalanced accuracy 90%, balanced accuracy 87%). The sensitivity for identifying PG was 94%, while the sensitivity forother chronic wound types was 97% for diabetic foot ulcers (DFU), 92% for venous leg ulcers (VLU), 78% for mixed leg ulcers and 74% for arterial leg ulcers.
Discussion
The machine learning model effectively differentiates PG from the most common leg and foot ulcers and was very accurate for classifying DFU and VLU. A higher rate of misclassifications occurred for the other vascular ulcers, that is, mixed and arterial leg ulcers. This aligns with the challenges in clinical practice.
Conclusion
Despite the limited number of wound images, this novel multiclass wound classification model accurately identified PG and differentiated leg and foot ulcer subtypes, providing a foundation for a diagnostic support system
Einfluss der Reduktion von Sojaextraktionsschrot im Ferkelaufzuchtfutter auf Wachstums- und Leistungsparameter
No full text
Pore Structure Analysis of Growing Media Using X-Ray Microcomputed Tomography
No full textThis study investigated the representative elementary volume (REV) for visible porosity in horticultural growing media (peat, commercial mixture, treated wood fibre/peat, pure wood fibre) using x-ray micro-computed tomography (µCT) with 2D and 3D image division, pore morphology, water retention curve (WRC), and saturated hydraulic conductivity (Ksat) via pore network modelling (PNM). Two sample sizes (10 × 10 cm, 3 × 3 cm, height × diameter) with resolutions of 46 and 15 µm were analysed. REV was assessed using deterministic (dREV) and statistical (sREV) criteria, evaluating the porosity and coefficient of variation across subvolumes. The results showed that 3D division of large samples achieved REV only for pure wood fibre (8000–10,000 µm), while 2D division met both criteria for all media. For small samples, 3D division achieved REV only for wood fibre/peat mixture, but 2D division succeeded for all media above 3000 µm. Pore analyses indicated that pure wood fibre had the largest, most connected pores, enhancing drainage, while peat showed complex, retentive structures. WRCs aligned well with lab data (R2 0.88). PNM Ksat estimates from small images were more accurate, with discrepancies (21–172%) due to segmentation artefacts. Future studies should incorporate permeability or tortuosity and explore multiscale imaging for improved hydrophysical predictions. This study also highlights advantages unique to X-ray µCT compared to standard laboratory methods, e.g., direct three-dimensional quantification of pore structure parameters and an image-based determination of the REV
Evaluation of a centrifuge quick test to determine the water retention parameters of turf rootzone materials
No full textThe common water retention parameters of materials used for rootzones of turfgrass are field capacity, air capacity, wilting point, and available water capacity. More measurement points are necessary to describe the entire water retention curve (WRC) correctly, for example, for model applications. Standard laboratory methods include sand suction tables and pressure plates, or technically more advanced methods using tensiometers and evaporation. All these methods require special equipment and are time‐consuming (at least 1 week); using the centrifuge method as a quick test to determine the whole WRC takes not more than 1 day. We compared water retention parameters and complete WRC of seven different root zone construction materials with various textures and organic matter contents determined by standard and centrifuge methods. Nearly all centrifuge results were within the reproducibility range of the standard method, thus demonstrating the application potential of this quick method
Comparative analysis of solar thermal and PV-assisted heat pump technologies for industrial process heat in Southern Africa and Central Europe
No full textIndustry is the largest energy consumer, with most demand occurring as direct heat input. For low to medium-temperature industrial processes (≤160°C), heat pumps and solar technologies offer significant opportunities to reduce industrial carbon emissions. This work presents a comprehensive comparison of solar thermal energy systems and photovoltaic-assisted heat pump technologies for industrial process heat applications, specifically contrasting their performance in Southern Africa and Central Europe.
Three distinct scenarios from the beverage industry, each representing different operational scales, are examined to highlight the trade-off between economic viability and environmental sustainability. Hybrid solar thermal configurations achieve high solar fractions, but the heat pump pathway proves very viable in both regions for deep, cost-effective CO2 reduction. While Southern African solar systems yield substantially higher energy output per square meter of collector area, the relatively higher cost of natural gas in Central Europe provides favorable economic returns. Significant CO2 reductions are enabled by a combination of measures, including public funding to offset initial capital costs, rising fossil energy prices, and decreasing technology costs
Quantifying feather pecking motivation in dual purpose chickens using a novel measurement device
No full textThis study developed a novel device to quantify feather pecking motivation in dual-purpose chickens, offering a standardized and continuous approach to behavioral assessment in poultry farming. Using a force gauge, the device measures pecking intensity in Newtons (N), providing objective, quantifiable insights into pecking motivation during the rearing phase. Despite current limitations in correlating observed and measured pecking events, ongoing refinements and additional testing will help to fully explore the potential of this system especially in advancing research on feather pecking mitigation strategies. The device bridges the gap between behavioral observations and quantifiable welfare metrics, driving innovation in poultry welfare research
Linking laying performance and individual animal welfare using RFID in poultry nests
No full textIn poultry farming, collecting individual data is challenging, particularly for laying performance. This study uses RFID (radio-frequency identification)-equipped electronic nests to correlate egg number, egg weight, and body weight with the welfare status of individual hens. Additionally, behavioral data from nest use and morphological scores, particularly keel bone health, are considered. The goal is to identify high-performing hens in flocks of local chicken breeds with good welfare status, improving animal management and breeding strategies through accurate, individualized monitoring
