Publications - Chair of Computer Science II

Journal and Magazine Articles

AI-Based Wound Measurement in a Telemedicine-Style Workflow: Agreement With In-Person Reference Measurements Compared With Individual Physicians and Expert Consensus. Dege, Tassilo; Borst, Vanessa; Deutzmann, Luisa; Maiwald, Jan Hendrik; Kounev, Samuel; Schmieder, Astrid; in International Wound Journal (2026).
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
- [ DOI ]
Background: Accurate wound size measurement is essential for monitoring healing, yet visual estimates are variable---especially in telemedicine-style assessments. Objective: To compare AI-based wound size estimates with physician image-based estimates using in-person measurements as ground truth (GT). Methods: In a prospective dataset of bedside wound photographs (Dec 2024--Apr 2025), GT height and width were measured in person by two experienced raters using a calibrated ruler. A set of ArUco marker placed adjacent to the wound at the same skin level enabled image-based retrieval of real-world size information. Five physicians estimated height and width from images in a telemedicine-style workflow. Two deep-learning segmentation architectures (TransNeXt-Tiny [TransNeXt] and VWFormerConvNeXtS [VW-Conv]) generated automated size estimates. Agreement and error were quantified using ICC(2,1), Pearson's r, MAE, bias, and Bland--Altman limits of agreement. Results: AI achieved excellent agreement for height (TransNeXt ICC = 0.96, MAE = 0.82 cm; VW-Conv ICC = 0.95, MAE = 0.68 cm) with minimal bias. For width, AI showed moderate-to-strong agreement with a tendency to underestimation (TransNeXt ICC = 0.86, MAE = 0.69 cm, bias = -0.42 cm; VW-Conv ICC = 0.77, MAE = 0.74 cm, bias = -0.64 cm). Individual physician performance was heterogeneous, while a human consensus improved agreement (height ICC = 0.94, MAE = 0.82 cm; width ICC = 0.91, MAE = 0.65 cm). Conclusion: AI-based wound measurement was more consistent than single physician estimates and competitive with a consensus of experts, supporting its use as an objective tool for standardized wound monitoring in telemedicine and clinical research.

@article{dege_2026_AIWoundMeasurement, abstract = {Background: Accurate wound size measurement is essential for monitoring healing, yet visual estimates are variable—especially in telemedicine-style assessments. Objective: To compare AI-based wound size estimates with physician image-based estimates using in-person measurements as ground truth (GT). Methods: In a prospective dataset of bedside wound photographs (Dec 2024–Apr 2025), GT height and width were measured in person by two experienced raters using a calibrated ruler. A set of ArUco marker placed adjacent to the wound at the same skin level enabled image-based retrieval of real-world size information. Five physicians estimated height and width from images in a telemedicine-style workflow. Two deep-learning segmentation architectures (TransNeXt-Tiny [TransNeXt] and VWFormerConvNeXtS [VW-Conv]) generated automated size estimates. Agreement and error were quantified using ICC(2,1), Pearson's r, MAE, bias, and Bland–Altman limits of agreement. Results: AI achieved excellent agreement for height (TransNeXt ICC = 0.96, MAE = 0.82 cm; VW-Conv ICC = 0.95, MAE = 0.68 cm) with minimal bias. For width, AI showed moderate-to-strong agreement with a tendency to underestimation (TransNeXt ICC = 0.86, MAE = 0.69 cm, bias = -0.42 cm; VW-Conv ICC = 0.77, MAE = 0.74 cm, bias = -0.64 cm). Individual physician performance was heterogeneous, while a human consensus improved agreement (height ICC = 0.94, MAE = 0.82 cm; width ICC = 0.91, MAE = 0.65 cm). Conclusion: AI-based wound measurement was more consistent than single physician estimates and competitive with a consensus of experts, supporting its use as an objective tool for standardized wound monitoring in telemedicine and clinical research.}, author = {Dege, Tassilo and Borst, Vanessa and Deutzmann, Luisa and Maiwald, Jan Hendrik and Kounev, Samuel and Schmieder, Astrid}, journal = {International Wound Journal}, keywords = {t_journalmagazine}, title = {{AI}-Based Wound Measurement in a Telemedicine-Style Workflow: Agreement With In-Person Reference Measurements Compared With Individual Physicians and Expert Consensus}, year = 2026 }
Patient-Centered Lupus Erythematosus Mobile Apps: Systematic Search and Cross-Sectional Evaluation by Patients and Physicians. Dege, Tassilo; Ullmann, Antonia; Glatzel, Caroline; Fleißner, Janik; Borst, Vanessa; Strunz, Patrick-Pascal; Schmalzing, Marc; Goebeler, Matthias; Schmieder, Astrid; in JMIR mHealth and uHealth (2026). 14 e73019.
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
- [ DOI ]
Background: Lupus erythematosus (LE) is a chronic autoimmune disease that significantly impacts patients' quality of life. Photosensitivity is a key impairment that severely limits the quality of life, especially in cutaneous lupus erythematosus (CLE), where exposure to sunlight can lead to rashes, exacerbations, and pain. In systemic lupus erythematosus (SLE), other manifestations such as joint pain, fatigue, and organ damage may contribute to decreased physical function and emotional distress. Mobile health apps (MHA) offer potential support for comprehensive disease management for the symptoms mentioned above. However, there is a lack of systematic analysis of available lupus management apps. Objective: This study aims to systematically identify publicly available German or English MHA for lupus management as well as to assess their quality by surveying both patients and physicians. Methods: A systematic search and assessment of German or English mobile apps for patients with lupus, available in the Google Play Store and Apple App Store, was conducted independently by two reviewers. The two apps that met all relevant criteria were then reviewed independently by seven physicians using the German Mobile Application Rating Scale (MARS) and the System Usability Scale (SUS). Subsequently, they were reviewed by five patients (three with SLE and two with CLE), using the user version of MARS (uMARS) and SUS. Additionally, the Affinity for Technology Interaction (ATI) scale was collected from both patients and physicians to evaluate the technical affinity in both groups. Results: In total, 29 apps were available on the Apple Store and 26 on the Google Store, with 18 apps being present and downloadable on both platforms. Of the 18 apps, 16 were excluded because they did not meet the inclusion and exclusion criteria. Only two apps, Lupus Log and Lupus Minder met all the required criteria and were included in the study. The mean MARS scores varied from 2.61/5 to 4.17/5 and mean SUS from 17.5/100 to 100/100 between physicians. The app with the highest mean overall MARS score was Lupus Log, which was rated with 3.91/5 on average by the physicians. Patients evaluated the app with a comparably mean uMARS score (3.95/5). Technical affinity, objectified by ATI, was higher in patients than physicians (3.9 vs 3.68). Conclusions: Systematic identification and evaluation showed high-quality apps for patient-centered lupus MHA as indicated by MARS and uMARS scores greater than 3 for both Lupus Log and Lupus Minder.

@article{dege_2026_lupus, abstract = {Background: Lupus erythematosus (LE) is a chronic autoimmune disease that significantly impacts patients' quality of life. Photosensitivity is a key impairment that severely limits the quality of life, especially in cutaneous lupus erythematosus (CLE), where exposure to sunlight can lead to rashes, exacerbations, and pain. In systemic lupus erythematosus (SLE), other manifestations such as joint pain, fatigue, and organ damage may contribute to decreased physical function and emotional distress. Mobile health apps (MHA) offer potential support for comprehensive disease management for the symptoms mentioned above. However, there is a lack of systematic analysis of available lupus management apps. Objective: This study aims to systematically identify publicly available German or English MHA for lupus management as well as to assess their quality by surveying both patients and physicians. Methods: A systematic search and assessment of German or English mobile apps for patients with lupus, available in the Google Play Store and Apple App Store, was conducted independently by two reviewers. The two apps that met all relevant criteria were then reviewed independently by seven physicians using the German Mobile Application Rating Scale (MARS) and the System Usability Scale (SUS). Subsequently, they were reviewed by five patients (three with SLE and two with CLE), using the user version of MARS (uMARS) and SUS. Additionally, the Affinity for Technology Interaction (ATI) scale was collected from both patients and physicians to evaluate the technical affinity in both groups. Results: In total, 29 apps were available on the Apple Store and 26 on the Google Store, with 18 apps being present and downloadable on both platforms. Of the 18 apps, 16 were excluded because they did not meet the inclusion and exclusion criteria. Only two apps, Lupus Log and Lupus Minder met all the required criteria and were included in the study. The mean MARS scores varied from 2.61/5 to 4.17/5 and mean SUS from 17.5/100 to 100/100 between physicians. The app with the highest mean overall MARS score was Lupus Log, which was rated with 3.91/5 on average by the physicians. Patients evaluated the app with a comparably mean uMARS score (3.95/5). Technical affinity, objectified by ATI, was higher in patients than physicians (3.9 vs 3.68). Conclusions: Systematic identification and evaluation showed high-quality apps for patient-centered lupus MHA as indicated by MARS and uMARS scores greater than 3 for both Lupus Log and Lupus Minder.}, author = {Dege, Tassilo and Ullmann, Antonia and Glatzel, Caroline and Flei{\ss}ner, Janik and Borst, Vanessa and Strunz, Patrick-Pascal and Schmalzing, Marc and Goebeler, Matthias and Schmieder, Astrid}, journal = {JMIR mHealth and uHealth}, keywords = {t_journalmagazine}, pages = {e73019}, title = {Patient-Centered Lupus Erythematosus Mobile Apps: Systematic Search and Cross-Sectional Evaluation by Patients and Physicians}, volume = 14, year = 2026 }
WoundAIssist: Development and Evaluation of an AI-Based Mobile Application for Remote Chronic Wound Care in Elderly Patients. Borst, Vanessa; Riedmann, Anna; Dege, Tassilo; Müller, Konstantin; Schmieder, Astrid; Lugrin, Birgit; Kounev, Samuel; in ACM Transactions on Computing for Healthcare (2026). Association for Computing Machinery.
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
- [ DOI ]
The rising prevalence of chronic wounds, especially in aging populations, presents a significant healthcare challenge due to prolonged hospitalizations, elevated costs, and reduced patient quality of life. Traditional wound care is resource-intensive, requiring frequent in-person visits that strain both patients and healthcare professionals (HCPs). Thus, we present WoundAIssist, a patient-centered, AI-driven mobile application supporting telemedical wound care. WoundAIssist enables patients to document wounds at home via photographs and questionnaires, while physicians remain engaged in the care process through remote monitoring and video consultations. A distinguishing feature is an integrated lightweight deep learning model for on-device wound segmentation, guiding users during image capture. Combined with patient-reported data and server-side AI analysis, this enables continuous monitoring of wound healing progression. Developed through an iterative, user-centered process involving patients and domain experts, WoundAIssist prioritizes an user-friendly design, particularly for elderly patients. A conclusive usability study with patients and dermatologists reported excellent usability, good app quality, and favorable perceptions of the AI-driven wound recognition. Our main contribution is two-fold: (I) the development and (II) evaluation of WoundAIssist, an easy-to-use yet comprehensive telehealth solution designed to bridge the gap between patients and HCPs. Additionally, we synthesize design insights for remote patient monitoring apps, derived from over three years of interdisciplinary research, that may inform the development of similar digital health tools across clinical domains.

@article{borst2026woundAIssistdescartes, abstract = {The rising prevalence of chronic wounds, especially in aging populations, presents a significant healthcare challenge due to prolonged hospitalizations, elevated costs, and reduced patient quality of life. Traditional wound care is resource-intensive, requiring frequent in-person visits that strain both patients and healthcare professionals (HCPs). Thus, we present WoundAIssist, a patient-centered, AI-driven mobile application supporting telemedical wound care. WoundAIssist enables patients to document wounds at home via photographs and questionnaires, while physicians remain engaged in the care process through remote monitoring and video consultations. A distinguishing feature is an integrated lightweight deep learning model for on-device wound segmentation, guiding users during image capture. Combined with patient-reported data and server-side AI analysis, this enables continuous monitoring of wound healing progression. Developed through an iterative, user-centered process involving patients and domain experts, WoundAIssist prioritizes an user-friendly design, particularly for elderly patients. A conclusive usability study with patients and dermatologists reported excellent usability, good app quality, and favorable perceptions of the AI-driven wound recognition. Our main contribution is two-fold: (I) the development and (II) evaluation of WoundAIssist, an easy-to-use yet comprehensive telehealth solution designed to bridge the gap between patients and HCPs. Additionally, we synthesize design insights for remote patient monitoring apps, derived from over three years of interdisciplinary research, that may inform the development of similar digital health tools across clinical domains.}, author = {Borst, Vanessa and Riedmann, Anna and Dege, Tassilo and M\"uller, Konstantin and Schmieder, Astrid and Lugrin, Birgit and Kounev, Samuel}, journal = {ACM Transactions on Computing for Healthcare}, keywords = {t_journalmagazine}, publisher = {Association for Computing Machinery}, title = {{WoundAIssist}: Development and Evaluation of an {AI}-Based Mobile Application for Remote Chronic Wound Care in Elderly Patients}, year = 2026 }
Patient-Centered Chronic Wound Care Mobile Apps: Systematic Identification, Analysis, and Assessment. Dege, Tassilo; Glatzel, Bernadette; Borst, Vanessa; Grän, Franziska; Goller, Simon; Glatzel, Caroline; Goebeler, Matthias; Schmieder, Astrid; in JMIR mHealth and uHealth (2024). 12
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
- [ DOI ]
Background: The prevalence of chronic wounds is predicted to increase within the aging populations in industrialized countries. Patients experience significant distress due to pain, wound secretions, and the resulting immobilization. As the number of wounds continues to rise, their adequate care becomes increasingly costly in terms of health care resources worldwide. eHealth support systems are being increasingly integrated into patient care. However, to date, no systematic analysis of such apps for chronic wounds has been published. Objective: The aims of this study were to systematically identify and subjectively assess publicly available German- or English-language mobile apps for patients with chronic wounds, with quality assessments performed by both patients and physicians. Methods: Two reviewers independently conducted a systematic search and assessment of German- or English-language mobile apps for patients with chronic wounds that were available in the Google Play Store and Apple App Store from April 2022 to May 2022. In total, 3 apps met the inclusion and exclusion criteria and were reviewed independently by 10 physicians using the German Mobile App Rating Scale (MARS) and the System Usability Scale (SUS). The app with the highest mean MARS score was subsequently reviewed by 11 patients with chronic wounds using the German user version of the MARS (uMARS) and the SUS. Additionally, Affinity for Technology Interaction (ATI) scale scores were collected from both patients and physicians. Results: This study assessed mobile apps for patients with chronic wounds that were selected from a pool of 118 identified apps. Of the 73 apps available in both app stores, 10 were patient oriented. After excluding apps with advertisements or costs, 3 apps were evaluated by 10 physicians. Mean MARS scores ranged from 2.64 (SD 0.65) to 3.88 (SD 0.65) out of 5, and mean SUS scores ranged from 50.75 (SD 27) to 80.5 (SD 17.7) out of 100. WUND APP received the highest mean MARS score (mean 3.88, SD 0.65 out of 5) among physicians. Hence, it was subsequently assessed by 11 patients and achieved a similar rating (uMARS score: mean 3.89, SD 0.4 out of 5). Technical affinity, as measured with the ATI scale, was slightly lower in patients (score: mean 3.62, SD 1.35 out of 6) compared to physicians (score: mean 3.88, SD 1.03 out 6). Conclusions: The quality ratings from physicians and patients were comparable and indicated mediocre app quality. Technical affinity, as assessed by using the ATI scale, was slightly lower for patients. Adequate apps for patients with chronic wounds remain limited, emphasizing the need for improved app development to meet patient needs. The ATI scale proved valuable for assessing technical affinity among different user groups.

@article{dege2024patientcentered, abstract = {Background: The prevalence of chronic wounds is predicted to increase within the aging populations in industrialized countries. Patients experience significant distress due to pain, wound secretions, and the resulting immobilization. As the number of wounds continues to rise, their adequate care becomes increasingly costly in terms of health care resources worldwide. eHealth support systems are being increasingly integrated into patient care. However, to date, no systematic analysis of such apps for chronic wounds has been published. Objective: The aims of this study were to systematically identify and subjectively assess publicly available German- or English-language mobile apps for patients with chronic wounds, with quality assessments performed by both patients and physicians. Methods: Two reviewers independently conducted a systematic search and assessment of German- or English-language mobile apps for patients with chronic wounds that were available in the Google Play Store and Apple App Store from April 2022 to May 2022. In total, 3 apps met the inclusion and exclusion criteria and were reviewed independently by 10 physicians using the German Mobile App Rating Scale (MARS) and the System Usability Scale (SUS). The app with the highest mean MARS score was subsequently reviewed by 11 patients with chronic wounds using the German user version of the MARS (uMARS) and the SUS. Additionally, Affinity for Technology Interaction (ATI) scale scores were collected from both patients and physicians. Results: This study assessed mobile apps for patients with chronic wounds that were selected from a pool of 118 identified apps. Of the 73 apps available in both app stores, 10 were patient oriented. After excluding apps with advertisements or costs, 3 apps were evaluated by 10 physicians. Mean MARS scores ranged from 2.64 (SD 0.65) to 3.88 (SD 0.65) out of 5, and mean SUS scores ranged from 50.75 (SD 27) to 80.5 (SD 17.7) out of 100. WUND APP received the highest mean MARS score (mean 3.88, SD 0.65 out of 5) among physicians. Hence, it was subsequently assessed by 11 patients and achieved a similar rating (uMARS score: mean 3.89, SD 0.4 out of 5). Technical affinity, as measured with the ATI scale, was slightly lower in patients (score: mean 3.62, SD 1.35 out of 6) compared to physicians (score: mean 3.88, SD 1.03 out 6). Conclusions: The quality ratings from physicians and patients were comparable and indicated mediocre app quality. Technical affinity, as assessed by using the ATI scale, was slightly lower for patients. Adequate apps for patients with chronic wounds remain limited, emphasizing the need for improved app development to meet patient needs. The ATI scale proved valuable for assessing technical affinity among different user groups.}, author = {Dege, Tassilo and Glatzel, Bernadette and Borst, Vanessa and Grän, Franziska and Goller, Simon and Glatzel, Caroline and Goebeler, Matthias and Schmieder, Astrid}, journal = {JMIR mHealth and uHealth}, keywords = {t_journalmagazine}, month = {03}, title = {Patient-Centered Chronic Wound Care Mobile Apps: Systematic Identification, Analysis, and Assessment}, volume = 12, year = 2024 }
Deep Neural Network Regression for Normalized Digital Surface Model Generation with Sentinel-2 Imagery. Müller, Konstantin; Leppich, Robert; Geiß, Christian; Borst, Vanessa; Pelizari, Patrick Aravena; Kounev, Samuel; in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (2023). 1–14.
- [ Abstract ]
- [ BibTeX ]
- [ DOI ]
In recent history, normalized digital surface models (nDSMs) have been constantly gaining importance as a means to solve large-scale geographic problems. High-resolution surface models are precious, as they can provide detailed information for a specific area. However, measurements with a high resolution are time-consuming and costly. Only a few approaches exist to create high-resolution normalized digital surface models for extensive areas. This paper explores approaches to extract high-resolution nDSMs from low-resolution Sentinel-2 data, allowing us to derive large-scale models. We thereby utilize the advantages of Sentinel 2 being open access, having global coverage, and providing steady updates through a high repetition rate. Several deep-learning models are trained to overcome the gap in producing high-resolution surface maps from low-resolution input data. With U-Net as a base architecture, we extend the capabilities of our model by integrating tailored multiscale encoders with differently sized kernels in the convolution as well as conformed self-attention inside the skip connection gates. Using pixel-wise regression, our U-Net base models can achieve a mean height error of approximately two meters. Moreover, through our enhancements to the model architecture, we reduce the model error by more than seven percent.

@article{10189905, abstract = {In recent history, normalized digital surface models (nDSMs) have been constantly gaining importance as a means to solve large-scale geographic problems. High-resolution surface models are precious, as they can provide detailed information for a specific area. However, measurements with a high resolution are time-consuming and costly. Only a few approaches exist to create high-resolution normalized digital surface models for extensive areas. This paper explores approaches to extract high-resolution nDSMs from low-resolution Sentinel-2 data, allowing us to derive large-scale models. We thereby utilize the advantages of Sentinel 2 being open access, having global coverage, and providing steady updates through a high repetition rate. Several deep-learning models are trained to overcome the gap in producing high-resolution surface maps from low-resolution input data. With U-Net as a base architecture, we extend the capabilities of our model by integrating tailored multiscale encoders with differently sized kernels in the convolution as well as conformed self-attention inside the skip connection gates. Using pixel-wise regression, our U-Net base models can achieve a mean height error of approximately two meters. Moreover, through our enhancements to the model architecture, we reduce the model error by more than seven percent.}, author = {Müller, Konstantin and Leppich, Robert and Geiß, Christian and Borst, Vanessa and Pelizari, Patrick Aravena and Kounev, Samuel}, journal = {IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, keywords = {t_journalmagazine}, pages = {1-14}, title = {Deep Neural Network Regression for Normalized Digital Surface Model Generation with Sentinel-2 Imagery}, year = 2023 }

Full Papers

MedSegBenchmarker: A Raw-Count-First Framework for Controlled 2D Medical Image Segmentation Benchmarks. Borst, Vanessa; Horn, Lukas; Grillmeyer, Daniel; Prantl, Thomas; Kounev, Samuel; in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) (2027).

Just accepted.
- [ BibTeX ]
@inproceedings{Borst_2027_medsegbenchmarker, author = {Borst, Vanessa and Horn, Lukas and Grillmeyer, Daniel and Prantl, Thomas and Kounev, Samuel}, booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)}, keywords = {t_interdisciplinary}, note = {Just accepted.}, title = {{MedSegBenchmarker}: A Raw-Count-First Framework for Controlled {2D} Medical Image Segmentation Benchmarks}, year = 2027 }
Measuring Compression Honestly: A Dense-FP16 Baseline for Quantization and Sparsity on CNNs and Vision Transformers. Prantl, Thomas; Barthel, Til; Borst, Vanessa; Kounev, Samuel; in 2026 34rd International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems (MASCOTS) (2026). IEEE.

To appear.
- [ BibTeX ]
@inproceedings{Sardine_2027_Mascots_AI_Compression, author = {Prantl, Thomas and Barthel, Til and Borst, Vanessa and Kounev, Samuel}, booktitle = {2026 34rd International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems (MASCOTS)}, keywords = {descartes}, month = 10, note = {To appear.}, publisher = {IEEE}, title = {Measuring Compression Honestly: A Dense-FP16 Baseline for Quantization and Sparsity on CNNs and Vision Transformers}, year = 2026 }
GP-VM×SMA: Benchmarking General-Purpose Vision Models and Specialized Architectures for 2D Medical Image Segmentation. Borst, Vanessa; Riedmann, Anna; Kounev, Samuel; in 37th British Machine Vision Conference (BMVC) (2026). BMVA.
- [ Abstract ]
- [ BibTeX ]
Medical image segmentation (MIS) is a fundamental component of computer-assisted diagnosis and clinical decision support. Over the past decade, numerous architectures specifically tailored to medical imaging have emerged to address domain-specific challenges such as low contrast, small anatomical structures, and limited annotated data. In parallel, rapid progress in computer vision has produced highly capable general-purpose vision models (GP-VMs) originally designed for natural images. Despite their strong performance on standard vision benchmarks, their effectiveness for MIS remains insufficiently understood. In this controlled empirical study, we examine whether specialized medical segmentation architectures (SMAs) provide systematic advantages over modern GP-VMs for 2D MIS. We compare eleven SMAs and GP-VMs using a unified training and evaluation protocol across three heterogeneous datasets, each covering different 2D imaging modalities, class structures, and data characteristics. Beyond segmentation performance, we employ linear mixed-effects models (LMMs) for statistical assessment and qualitative Grad-CAM visualizations to investigate explainability (XAI)-related model behavior. Our results imply that, for the analyzed settings, GP-VMs achieve performance comparable to that of specialized MIS models. Moreover, XAI analyses indicate that GP-VMs are capable of identifying clinically relevant structures despite the absence of explicit domain-specific architectural priors. These findings suggest that GP-VMs can represent a viable alternative to domain-specific methods for 2D MIS, highlighting the importance of informed model selection. All code and resources are available at GitHub.

@inproceedings{Borst_2026_GPVMxSMA, abstract = {Medical image segmentation (MIS) is a fundamental component of computer-assisted diagnosis and clinical decision support. Over the past decade, numerous architectures specifically tailored to medical imaging have emerged to address domain-specific challenges such as low contrast, small anatomical structures, and limited annotated data. In parallel, rapid progress in computer vision has produced highly capable general-purpose vision models (GP-VMs) originally designed for natural images. Despite their strong performance on standard vision benchmarks, their effectiveness for MIS remains insufficiently understood. In this controlled empirical study, we examine whether specialized medical segmentation architectures (SMAs) provide systematic advantages over modern GP-VMs for 2D MIS. We compare eleven SMAs and GP-VMs using a unified training and evaluation protocol across three heterogeneous datasets, each covering different 2D imaging modalities, class structures, and data characteristics. Beyond segmentation performance, we employ linear mixed-effects models (LMMs) for statistical assessment and qualitative Grad-CAM visualizations to investigate explainability (XAI)-related model behavior. Our results imply that, for the analyzed settings, GP-VMs achieve performance comparable to that of specialized MIS models. Moreover, XAI analyses indicate that GP-VMs are capable of identifying clinically relevant structures despite the absence of explicit domain-specific architectural priors. These findings suggest that GP-VMs can represent a viable alternative to domain-specific methods for 2D MIS, highlighting the importance of informed model selection. All code and resources are available at GitHub.}, author = {Borst, Vanessa and Riedmann, Anna and Kounev, Samuel}, booktitle = {37th British Machine Vision Conference (BMVC)}, keywords = {t_interdisciplinary}, publisher = {BMVA}, title = {{{GP-VM×SMA}: Benchmarking General-Purpose Vision Models and Specialized Architectures for {2D} Medical Image Segmentation}}, year = 2026 }
WoundAmbit: Bridging State-of-the-Art Semantic Segmentation and Real-World Wound Care. Borst, Vanessa; Dittus, Timo; Dege, Tassilo; Schmieder, Astrid; Kounev, Samuel; in Proceedings of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD). Applied Data Science Track (2025). Springer Nature Switzerland, Porto, Portugal.

Best Student Paper ADS Track
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
- [ DOI ]
Chronic wounds affect a large population, particularly the elderly and diabetic patients, who often exhibit limited mobility and co-existing health conditions. Automated wound monitoring via mobile image capture can reduce in-person physician visits by enabling remote tracking of wound size. Semantic segmentation is key to this process, yet wound segmentation remains underrepresented in medical imaging research. To address this, we benchmark state-of-the-art deep learning models from general-purpose vision, medical imaging, and top methods from public wound challenges. For a fair comparison, we standardize training, data augmentation, and evaluation, conducting cross-validation to minimize partitioning bias. We also assess real-world deployment aspects, including generalization to an out-of-distribution wound dataset, computational efficiency, and interpretability. Additionally, we propose a reference object-based approach to convert AI-generated masks into clinically relevant wound size estimates and evaluate this, along with mask quality, for the five best architectures based on physician assessments. Overall, the transformer-based TransNeXt showed the highest levels of generalizability. Despite variations in inference times, all models processed at least one image per second on the CPU, which is deemed adequate for the intended application. Interpretability analysis typically revealed prominent activations in wound regions, emphasizing focus on clinically relevant features. Expert evaluation showed high mask approval for all analyzed models, with VWFormer and ConvNeXtS backbone performing the best. Size retrieval accuracy was similar across models, and predictions closely matched expert annotations. Finally, we demonstrate how our AI-driven wound size estimation framework, WoundAmbit, is integrated into a custom telehealth system. Our code and supplementary material are available on GitHub and Zenodo, respectively.

@inproceedings{borst2025woundambit, abstract = {Chronic wounds affect a large population, particularly the elderly and diabetic patients, who often exhibit limited mobility and co-existing health conditions. Automated wound monitoring via mobile image capture can reduce in-person physician visits by enabling remote tracking of wound size. Semantic segmentation is key to this process, yet wound segmentation remains underrepresented in medical imaging research. To address this, we benchmark state-of-the-art deep learning models from general-purpose vision, medical imaging, and top methods from public wound challenges. For a fair comparison, we standardize training, data augmentation, and evaluation, conducting cross-validation to minimize partitioning bias. We also assess real-world deployment aspects, including generalization to an out-of-distribution wound dataset, computational efficiency, and interpretability. Additionally, we propose a reference object-based approach to convert AI-generated masks into clinically relevant wound size estimates and evaluate this, along with mask quality, for the five best architectures based on physician assessments. Overall, the transformer-based TransNeXt showed the highest levels of generalizability. Despite variations in inference times, all models processed at least one image per second on the CPU, which is deemed adequate for the intended application. Interpretability analysis typically revealed prominent activations in wound regions, emphasizing focus on clinically relevant features. Expert evaluation showed high mask approval for all analyzed models, with VWFormer and ConvNeXtS backbone performing the best. Size retrieval accuracy was similar across models, and predictions closely matched expert annotations. Finally, we demonstrate how our AI-driven wound size estimation framework, WoundAmbit, is integrated into a custom telehealth system. Our code and supplementary material are available on GitHub and Zenodo, respectively.}, address = {Porto, Portugal}, author = {Borst, Vanessa and Dittus, Timo and Dege, Tassilo and Schmieder, Astrid and Kounev, Samuel}, booktitle = {Proceedings of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD). Applied Data Science Track}, keywords = {t_interdisciplinary}, month = {09}, note = {Best Student Paper ADS Track}, publisher = {Springer Nature Switzerland}, title = {{WoundAmbit}: Bridging State-of-the-Art Semantic Segmentation and Real-World Wound Care}, year = 2025 }
Privacy-Preserving Training and Inference of CNNs: A Medical Performance Case Study. Prantl, Thomas; Schneider, Louis; Engel, Simon; Horn, Lukas; Borst, Vanessa; Krupitzer, Christian; Kounev, Samuel; Bonilla, Rafael; in Proceedings of the 5th Intelligent Cybersecurity Conference (2025). 303–311. IEEE.
- [ BibTeX ]
- [ DOI ]
@inproceedings{MPC_Sardinen_2025, author = {Prantl, Thomas and Schneider, Louis and Engel, Simon and Horn, Lukas and Borst, Vanessa and Krupitzer, Christian and Kounev, Samuel and Bonilla, Rafael}, booktitle = {Proceedings of the 5th Intelligent Cybersecurity Conference}, keywords = {descartes}, month = {05}, pages = {303-311}, publisher = {IEEE}, series = {ICSC2025}, title = {Privacy-Preserving Training and Inference of CNNs: A Medical Performance Case Study}, year = 2025 }
Quantifying Data Leakage in Failure Prediction Tasks. Grillmeyer, Daniel; Hadry, Marius; Lesch, Veronika; Borst, Vanessa; Leppich, Robert; Bauer, André; Kounev, Samuel; in Proceedings of the 16th ACM/SPEC International Conference on Performance Engineering (2025). 334–346. Association for Computing Machinery, Toronto ON, Canada.
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
- [ DOI ]
With the ever increasing importance of cloud computing and a strong focus on reliable data centers, a high amount of research has been done on failure prediction for hard disk drives. The collection of monitoring data, such as SMART statistics (Self-Monitoring, Analysis, and Reporting Technology) from operational HDDs, enables operators to obtain predictions about the expected remaining useful life. Numerous methods for HDD failure prediction have been published in recent years, and their evaluation has shown decent results. However, a naive splitting into training and test sets can lead to data leakage and, thus, over-optimistic results that cannot be achieved on the data of scientific interest. In this paper, we propose a novel data leakage measure for quantifying the amount of data leakage in training and test datasets. Further, we define four splitting techniques and evaluate our measure in terms of the performance optimism of classification models with respect to these different splitting strategies. Our results consistently show that splitting techniques prone to data leakage induce an overestimation of predictive performance. Overall, we were able to show the usefulness of the defined data leakage measure, as well as its connection with different splitting techniques and the performance optimism of prediction models.

@inproceedings{10.1145/3676151.3719368, abstract = {With the ever increasing importance of cloud computing and a strong focus on reliable data centers, a high amount of research has been done on failure prediction for hard disk drives. The collection of monitoring data, such as SMART statistics (Self-Monitoring, Analysis, and Reporting Technology) from operational HDDs, enables operators to obtain predictions about the expected remaining useful life. Numerous methods for HDD failure prediction have been published in recent years, and their evaluation has shown decent results. However, a naive splitting into training and test sets can lead to data leakage and, thus, over-optimistic results that cannot be achieved on the data of scientific interest. In this paper, we propose a novel data leakage measure for quantifying the amount of data leakage in training and test datasets. Further, we define four splitting techniques and evaluate our measure in terms of the performance optimism of classification models with respect to these different splitting strategies. Our results consistently show that splitting techniques prone to data leakage induce an overestimation of predictive performance. Overall, we were able to show the usefulness of the defined data leakage measure, as well as its connection with different splitting techniques and the performance optimism of prediction models.}, address = {New York, NY, USA}, author = {Grillmeyer, Daniel and Hadry, Marius and Lesch, Veronika and Borst, Vanessa and Leppich, Robert and Bauer, Andr\'{e} and Kounev, Samuel}, booktitle = {Proceedings of the 16th ACM/SPEC International Conference on Performance Engineering}, keywords = {descartes}, pages = {334–346}, publisher = {Association for Computing Machinery}, series = {ICPE '25}, title = {Quantifying Data Leakage in Failure Prediction Tasks}, year = 2025 }
MACRO: A Multi-Head Attentional Convolutional Recurrent Network for the Classification of Co-Occurring Diseases in 12-Lead ECGs. Borst, Vanessa; Leppich, Robert; Kounev, Samuel; in Proceedings of the 2024 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI) (2024). 1–8. IEEE, Houston, TX.
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
- [ DOI ]
Cardiovascular diseases (CVDs) are a significant global health concern, causing more deaths than several types of cancer combined. Early detection and proper treatment are crucial for better clinical outcomes. Electrocardiograms (ECGs) offer valuable insights into the presence of CVDs. However, extracting powerful features from raw ECG signals for reliable automated diagnostics remains challenging due to high inter- and intra-patient variability, diversity of rhythmic and morphological abnormalities, and noise distributions. This paper proposes a deep learning architecture for the automated detection of diseases in 12-lead ECG data, capable of recognizing concurrent irregularities through multi-label classification. Moreover, we present a novel method that combines deep feature extraction with binary machine learning classifiers. To account for the distinct characteristics of various ECG leads, we use a multi-loss optimization strategy. Our methodology is rigorously evaluated through 10-fold cross-validation using the publicly available CPSC 2018 dataset. With macro F1 and AUC scores reaching up to 85.2% and 98.0%, respectively, our approach demonstrates advantage over existing state-of-the-art methods. At the same time, our architecture remains lightweight with approximately 1.7 million trainable parameters, which represents a reduction in the number of parameters of up to 68% compared to previous methods. Assessing the generalizability of our approach, we further evaluated it on the PTB-XL dataset and achieved macro F1, AUC, and accuracy scores comparable to existing methods, demonstrating the robustness of our model across diverse datasets. This advancement holds promise for enhanced automated diagnosis and improved patient care in the context of CVDs. Our code is available at: https://github.com/VanessaBorst/MACRO.

@inproceedings{borst2024MACRO, abstract = {Cardiovascular diseases (CVDs) are a significant global health concern, causing more deaths than several types of cancer combined. Early detection and proper treatment are crucial for better clinical outcomes. Electrocardiograms (ECGs) offer valuable insights into the presence of CVDs. However, extracting powerful features from raw ECG signals for reliable automated diagnostics remains challenging due to high inter- and intra-patient variability, diversity of rhythmic and morphological abnormalities, and noise distributions. This paper proposes a deep learning architecture for the automated detection of diseases in 12-lead ECG data, capable of recognizing concurrent irregularities through multi-label classification. Moreover, we present a novel method that combines deep feature extraction with binary machine learning classifiers. To account for the distinct characteristics of various ECG leads, we use a multi-loss optimization strategy. Our methodology is rigorously evaluated through 10-fold cross-validation using the publicly available CPSC 2018 dataset. With macro F1 and AUC scores reaching up to 85.2% and 98.0%, respectively, our approach demonstrates advantage over existing state-of-the-art methods. At the same time, our architecture remains lightweight with approximately 1.7 million trainable parameters, which represents a reduction in the number of parameters of up to 68% compared to previous methods. Assessing the generalizability of our approach, we further evaluated it on the PTB-XL dataset and achieved macro F1, AUC, and accuracy scores comparable to existing methods, demonstrating the robustness of our model across diverse datasets. This advancement holds promise for enhanced automated diagnosis and improved patient care in the context of CVDs. Our code is available at: https://github.com/VanessaBorst/MACRO.}, address = {Houston, TX}, author = {Borst, Vanessa and Leppich, Robert and Kounev, Samuel}, booktitle = {Proceedings of the 2024 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)}, keywords = {t_interdisciplinary}, organization = {IEEE}, pages = {1-8}, title = {{MACRO}: A Multi-Head Attentional Convolutional Recurrent Network for the Classification of Co-Occurring Diseases in 12-Lead {ECGs}}, year = 2024 }
Personal Task Design Preferences of Crowdworkers. Hirth, Matthias; Borchert, Kathrin; de Moor, Katrien; Borst, Vanessa; Hoßfeld, Tobias; in 12th International Conference on Quality of Multimedia Experience (QoMEX) (2020). Athlone, Ireland.
- [ BibTeX ]
@inproceedings{info3-inproceedings-2020-5, address = {Athlone, Ireland}, author = {Hirth, Matthias and Borchert, Kathrin and de Moor, Katrien and Borst, Vanessa and Hoßfeld, Tobias}, booktitle = {12th International Conference on Quality of Multimedia Experience (QoMEX)}, keywords = {t_full}, month = {05}, title = {Personal Task Design Preferences of Crowdworkers}, year = 2020 }

Short Papers

A Highly Customizable Platform Enabling Sophisticated Medical Eye-Tracking Studies. Bücheler, Vinzent; Dege, Tassilo; Pohl, Christofer; Hieber, Daniel; Borst, Vanessa; Pryss, Rüdiger; Schmieder, Astrid; Schobel, Johannes; in 39th International Symposium on Computer-Based Medical Systems (CBMS) (2026). 590–593.
- [ Abstract ]
- [ BibTeX ]
- [ DOI ]
Eye-tracking has demonstrated significant potential for analyzing visual attention in medical research. Nevertheless, technical challenges persist in establishing robust and customizable study designs ensuring reproducible data collection in complex scenarios. In prior work, we proposed a vendor-agnostic eye-tracking platform for conducting studies in digital (neuro) pathology. However, the initial platform was validated in an expert-supervised setting and lacked systematic evaluation of interaction design, user onboarding, and data quality mechanisms. Additionally, it exhibited a high degree of specificity, being closely associated with a particular medical domain. In order to address these limitations, a comprehensive set of cross-domain requirements was elicited and incorporated into a requirement-driven redesign resulting in a refined modular eye-tracking platform that was subsequently implemented and empirically validated. The redesigned platform contributes configurable multi-class task handling, flexible study flow components, and integrated ground truth--based feedback mechanisms. The platform was evaluated with 14 physicians from dermatology performing multi-class wound classification under eye-tracking conditions. Quantitative fixation analyses with respect to labeled regions of interest were combined with standardized usability assessment using the System Usability Scale (SUS). Findings indicate the system's dependable performance in a clinical setting and its exceptional usability, as evidenced by an average SUS score of 83.75. The findings of this study indicate that the refined platform enables standardized, reproducible, and domain-agnostic eye-tracking studies while significantly lowering technical barriers for study execution in clinical research environments.

@inproceedings{buecheler_2026_eyeTrackingPlatform, abstract = {Eye-tracking has demonstrated significant potential for analyzing visual attention in medical research. Nevertheless, technical challenges persist in establishing robust and customizable study designs ensuring reproducible data collection in complex scenarios. In prior work, we proposed a vendor-agnostic eye-tracking platform for conducting studies in digital (neuro) pathology. However, the initial platform was validated in an expert-supervised setting and lacked systematic evaluation of interaction design, user onboarding, and data quality mechanisms. Additionally, it exhibited a high degree of specificity, being closely associated with a particular medical domain. In order to address these limitations, a comprehensive set of cross-domain requirements was elicited and incorporated into a requirement-driven redesign resulting in a refined modular eye-tracking platform that was subsequently implemented and empirically validated. The redesigned platform contributes configurable multi-class task handling, flexible study flow components, and integrated ground truth–based feedback mechanisms. The platform was evaluated with 14 physicians from dermatology performing multi-class wound classification under eye-tracking conditions. Quantitative fixation analyses with respect to labeled regions of interest were combined with standardized usability assessment using the System Usability Scale (SUS). Findings indicate the system's dependable performance in a clinical setting and its exceptional usability, as evidenced by an average SUS score of 83.75. The findings of this study indicate that the refined platform enables standardized, reproducible, and domain-agnostic eye-tracking studies while significantly lowering technical barriers for study execution in clinical research environments.}, author = {B\"ucheler, Vinzent and Dege, Tassilo and Pohl, Christofer and Hieber, Daniel and Borst, Vanessa and Pryss, R\"udiger and Schmieder, Astrid and Schobel, Johannes}, booktitle = {39th International Symposium on Computer-Based Medical Systems (CBMS)}, keywords = {t_interdisciplinary}, pages = {590–593}, title = {A Highly Customizable Platform Enabling Sophisticated Medical Eye-Tracking Studies}, year = 2026 }
Early Explorations of Lightweight Models for Wound Segmentation on Mobile Devices. Borst, Vanessa; Dittus, Timo; Müller, Konstantin; Kounev, Samuel; in Proceedings of the German Conference on Artificial Intelligence (KI) (2024). Würzburg, Germany.
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
- [ DOI ]
- [ Download ]
The aging population poses numerous challenges to healthcare, including the increase in chronic wounds in the elderly. The current approach to wound assessment by therapists based on photographic documentation is subjective, highlighting the need for computer-aided wound recognition from smartphone photos. This offers objective and convenient therapy monitoring, while being accessible to patients from their home at any time. However, despite research in mobile image segmentation, there is a lack of focus on mobile wound segmentation. To address this gap, we conduct initial research on three lightweight architectures to investigate their suitability for smartphone-based wound segmentation. Using public datasets and UNet as a baseline, our results are promising, with both ENet and TopFormer, as well as the larger UNeXt variant, showing comparable performance to UNet. Furthermore, we deploy the models into a smartphone app for visual assessment of live segmentation, where results demonstrate the effectiveness of TopFormer in distinguishing wounds from wound-coloured objects. While our study highlights the potential of transformer models for mobile wound segmentation, future work should aim to further improve the mask contours. (An extended version of this paper is available for download from arXiv).

@inproceedings{borst2024early, abstract = {The aging population poses numerous challenges to healthcare, including the increase in chronic wounds in the elderly. The current approach to wound assessment by therapists based on photographic documentation is subjective, highlighting the need for computer-aided wound recognition from smartphone photos. This offers objective and convenient therapy monitoring, while being accessible to patients from their home at any time. However, despite research in mobile image segmentation, there is a lack of focus on mobile wound segmentation. To address this gap, we conduct initial research on three lightweight architectures to investigate their suitability for smartphone-based wound segmentation. Using public datasets and UNet as a baseline, our results are promising, with both ENet and TopFormer, as well as the larger UNeXt variant, showing comparable performance to UNet. Furthermore, we deploy the models into a smartphone app for visual assessment of live segmentation, where results demonstrate the effectiveness of TopFormer in distinguishing wounds from wound-coloured objects. While our study highlights the potential of transformer models for mobile wound segmentation, future work should aim to further improve the mask contours. (An extended version of this paper is available for download from arXiv).}, address = {Würzburg, Germany}, author = {Borst, Vanessa and Dittus, Timo and Müller, Konstantin and Kounev, Samuel}, booktitle = {Proceedings of the German Conference on Artificial Intelligence (KI)}, keywords = {t_interdisciplinary}, title = {Early Explorations of Lightweight Models for Wound Segmentation on Mobile Devices}, year = 2024 }

Posters

Explainable AI-assisted remote monitoring of chronic wounds in dermatology: prospective app-based feasibility study (WUNDERKINT). Dege, Tassilo; Deutzmann, Luisa; Borst, Vanessa; Kounev, Samuel; Schmieder, Astrid; in Abstract book of the European Academy of Dermatology and Venereology (EADV) Congress 2026 (2026).

Poster Abstract (No. ID-28)
- [ BibTeX ]
@incollection{dege_2026_wunderkint, author = {Dege, Tassilo and Deutzmann, Luisa and Borst, Vanessa and Kounev, Samuel and Schmieder, Astrid}, booktitle = {Abstract book of the European Academy of Dermatology and Venereology (EADV) Congress 2026}, keywords = {t_poster}, note = {Poster Abstract (No. ID-28)}, title = {Explainable {AI}-assisted remote monitoring of chronic wounds in dermatology: prospective app-based feasibility study {(WUNDERKINT)}}, year = 2026 }
Segmentierung und Verlaufskontrolle chronischer Wunden durch künstliche Intelligenz mithilfe einer mobilen App (WUNDERKINT Studie). Borst, Vanessa; Dege, Tassilo; Grän, Franziska; Hamrouni, Sandra; Michel, Daniel; Hager, Valerie; Goebeler, Matthias; Leppich, Robert; Kounev, Samuel; Schmieder, Astrid; in Allergologie. JDDG: Journal der Deutschen Dermatologischen Gesellschaft (2023). (Vol. 21) 28–177. John Wiley & Sons Ltd., Berlin, Germany.

Poster Abstract (No. P322, pp. 175-176)
- [ BibTeX ]
- [ URL ]
- [ DOI ]
- [ Download ]
@incollection{borst_ddg23_v2, address = {Berlin, Germany}, author = {Borst, Vanessa and Dege, Tassilo and Grän, Franziska and Hamrouni, Sandra and Michel, Daniel and Hager, Valerie and Goebeler, Matthias and Leppich, Robert and Kounev, Samuel and Schmieder, Astrid}, booktitle = {Allergologie. JDDG: Journal der Deutschen Dermatologischen Gesellschaft}, keywords = {t_poster}, month = {04}, note = {Poster Abstract (No. P322, pp. 175-176)}, number = {S1}, pages = {28-177}, publisher = {John Wiley & Sons Ltd.}, title = {Segmentierung und Verlaufskontrolle chronischer Wunden durch künstliche Intelligenz mithilfe einer mobilen App (WUNDERKINT Studie)}, volume = 21, year = 2023 }

Extended Abstracts

Computational Dermatology – An Automatic System for Visually Classifying Rare Wound Diseases. Hieber, Daniel; Dege, Tassilo; Borst, Vanessa; Liesche-Starnecker, Friederike; Schobel, Johannes; Pryss, Rüdiger; Glatzel, Caroline; Kramer, Frank; Schmieder, Astrid; Müller, Dominik; in Abstractband of the 70th Annual Meeting of the German Association for Medical Informatics, Biometry and Epidemiology (GMDS) (2025). Jena, Germany.
- [ BibTeX ]
- [ URL ]
- [ DOI ]
@incollection{hieber2025computational, address = {Jena, Germany}, author = {Hieber, Daniel and Dege, Tassilo and Borst, Vanessa and Liesche-Starnecker, Friederike and Schobel, Johannes and Pryss, Rüdiger and Glatzel, Caroline and Kramer, Frank and Schmieder, Astrid and Müller, Dominik}, booktitle = {Abstractband of the 70th Annual Meeting of the German Association for Medical Informatics, Biometry and Epidemiology (GMDS)}, keywords = {t_interdisciplinary}, title = {Computational Dermatology – An Automatic System for Visually Classifying Rare Wound Diseases}, year = 2025 }

Press Coverage

Hautkrankheit HS besser verstehen und Rückfälle verhindern. UKW, Pressestelle; (2026, September). einBLICK University magazine.

https://www.uni-wuerzburg.de/aktuelles/einblick/single/news/hautkrankheit-hs/
- [ BibTeX ]
- [ URL ]
@misc{pressestelleuniwurzburg2026HS, author = {UKW, Pressestelle}, keywords = {t_interdisciplinary}, month = {09}, note = {https://www.uni-wuerzburg.de/aktuelles/einblick/single/news/hautkrankheit-hs/}, publisher = {einBLICK University magazine}, title = {Hautkrankheit HS besser verstehen und Rückfälle verhindern}, year = 2026 }
KI-Ideen aus Würzburg überzeugen beim 3. Healthcare Hackathon. UKW, Pressestelle; (2026, April 7). einBlick University magazine, Würzburg, Germany.

https://www.uni-wuerzburg.de/aktuelles/einblick/single/news/healthcare-hackathon-2026/
- [ BibTeX ]
- [ URL ]
@misc{ukw2026kiideen, address = {Würzburg, Germany}, author = {{Pressestelle UKW}}, keywords = {t_interdisciplinary}, month = {04}, note = {https://www.uni-wuerzburg.de/aktuelles/einblick/single/news/healthcare-hackathon-2026/}, publisher = {einBlick University magazine}, title = {KI-Ideen aus Würzburg überzeugen beim 3. Healthcare Hackathon}, year = 2026 }
Kontinuität statt Momentaufnahme: App analysiert Wundverlauf Engelbrecht, Pia; in Fachmagazin Orthopädie Technik (2026).
- [ BibTeX ]
@periodical{engelbrecht_2026_kontinuitaet, author = {Engelbrecht, Pia}, journal = {Fachmagazin Orthopädie Technik}, keywords = {t_interdisciplinary}, number = {10/26}, title = {Kontinuität statt Momentaufnahme: App analysiert Wundverlauf}, year = 2026 }
Das IT-Girl Beyer, Nora; in Studienmagazin “whyte blue” (2026). YUMPU Publishing.

https://www.yumpu.com/de/document/read/71076237/whyte-blue-2026/35
- [ BibTeX ]
- [ URL ]
@periodical{beyer2026itgirl, author = {Beyer, Nora}, journal = {Studienmagazin “whyte blue”}, keywords = {t_interdisciplinary}, note = {https://www.yumpu.com/de/document/read/71076237/whyte-blue-2026/35}, number = 6, publisher = {YUMPU Publishing}, title = {Das IT-Girl}, year = 2026 }
Chronische Wunden mit App dauerhaft im Blick. UKW, Pressestelle; (2025, September 30). einBLICK University magazine, Würzburg, Germany.

https://www.uni-wuerzburg.de/aktuelles/einblick/single/news/preis-vanessa-borst/
- [ BibTeX ]
- [ URL ]
@misc{borst2025chronische, address = {Würzburg, Germany}, author = {{Pressestelle UKW}}, keywords = {t_interdisciplinary}, month = {09}, note = {https://www.uni-wuerzburg.de/aktuelles/einblick/single/news/preis-vanessa-borst/}, publisher = {einBLICK University magazine}, title = {Chronische Wunden mit App dauerhaft im Blick}, year = 2025 }
Großes Festkonzert des Universitätsbundes in der Neubaukirche mit Preisverleihung. .; (2022, November 21). Main-Post GmbH, Würzburg, Germany.

https://www.mainpost.de/ueberregional/journal/produktinformation/grosses-festkonzert-des-universitaetsbundes-in-der-neubaukirche-mit-preisverleihung-art-10974859
- [ BibTeX ]
- [ URL ]
@misc{wunderkintSonderforschungspreis2022mainpost, address = {Würzburg, Germany}, author = {.}, keywords = {borst_press}, month = 11, note = {https://www.mainpost.de/ueberregional/journal/produktinformation/grosses-festkonzert-des-universitaetsbundes-in-der-neubaukirche-mit-preisverleihung-art-10974859}, publisher = {Main-Post GmbH}, title = {Großes Festkonzert des Universitätsbundes in der Neubaukirche mit Preisverleihung}, year = 2022 }
Vanessa Borst gehört zu den IT-Frauen des Jahres. Pressestelle Uni Würzburg; (2025, Dezember 2). einBLICK University magazine, Würzburg, Germany.

https://www.uni-wuerzburg.de/aktuelles/einblick/single/news/vanessa-borst-it-preis/
- [ BibTeX ]
- [ URL ]
@misc{pressestelleuniwurzburg2025vanessa, address = {Würzburg, Germany}, author = {{Pressestelle Uni Würzburg}}, keywords = {borst_press}, month = 12, note = {https://www.uni-wuerzburg.de/aktuelles/einblick/single/news/vanessa-borst-it-preis/}, publisher = {einBLICK University magazine}, title = {Vanessa Borst gehört zu den IT-Frauen des Jahres}, year = 2025 }

Bildnachweise