Bernabé García, Gregorio

Person:
Bernabé García, Gregorio

Name

Bernabé García, Gregorio

publication.page.department

Universidad de Murcia. Departamento de Ingeniería y Tecnología de Computadores

Full item page

Search Results

Now showing 1 - 8 of 8

Restricted
Application of YOLOv8 and a model based on vision transformers and UNet for LVNC diagnosis: advantages and limitations
(Springer, 2025-04-25) Haro Orenes, Salvador de; González Férez, María Pilar; García Carrasco, José Manuel; Bernabé García, Gregorio; Ingeniería y Tecnología de Computadores
Hypertrabeculation or left ventricular non-compaction (LVNC) is a cardiac condition that has recently been recognized. While several methods exist for accurately measuring the trabeculae in the ventricle, there is still no consensus within the medical community regarding the optimal approach. In previous work, we introduced DL-LVTQ, a tool based on a UNet convolutional neural network designed to quantify the trabeculae in the left ventricle. In this paper, we present an expanded dataset that includes new patients affected by a cardiomyopathy known as Titin, necessitating the retraining of the models involved in our study on this updated dataset to accurately infer future patients with this condition. We also introduce ViTUNet, a hybrid architecture that aims to merge the benefits of UNet and Vision Transformers for precise segmentation of the left ventricle. Furthermore, we train a YOLOv8 model to detect the left ventricle and integrate it with the hybrid model to focus segmentation on a region of interest around the ventricle. Regarding the precision quality achieved by ViTUNet using YOLOv8, results are quite similar to those obtained by the DL-LVTQ tool, suggesting that the dataset is a limiting factor in our improvement. To substantiate this, we conduct a detailed analysis of the MRI slices in the current dataset. By identifying and removing problematic slices, results significantly improve. The introduction of a YOLOv8 model alongside a deep learning model presents a promising approach.
Open Access
Survival risk prediction in hematopoietic stem cell transplantation for multiple myeloma
(De Gruyter, 2025-06-03) Belmonte, José María; Blanquer Blanquer, Miguel; Bernabé García, Gregorio; Jiménez Barrionuevo, Fernando; García Carrasco, José Manuel; Ingeniería y Tecnología de Computadores
This paper investigates the application of Survival Analysis (SA) techniques to forecast outcomes after autologous Hematopoietic Stem Cell Transplantation (aHSCT) for Multiple Myeloma (MM). By leveraging six SA models, we examine their predictive capabilities, measured through the Concordance Index (C-index) metric. Beyond evaluating model performance, we analyze feature importance using permutation and SHAP methods, highlighting key clinical factors such as treatment history, disease stage, and prior disease progression or relapse as critical predictors of survival. The findings suggest that while all models performed well based on the C-index, a detailed examination revealed variations in how each model processed data. Specifically, the Coxnet and Random Survival Forest models exhibited a more thorough use of clinical variables, whereas the gradient boosting models appeared to rely on a narrower range of features, potentially limiting their ability to differentiate between patients with comparable profiles. Risk predictions categorized patients into low, moderate, and high-risk levels. For lower-risk patients, the procedure showed positive outcomes, while higher-risk individuals were predicted to have limited survival benefits, recommending alternative treatments. Lastly, we propose future research to expand these models into time-to-event estimations, offering additional support for decision-making by predicting patient life expectancy post-transplant, considering their pre-transplant clinical attributes.
Restricted
Survival analysis in hematopoietic stem cell transplantation for multiple myeloma: methodology and survival predictions
(Springer Nature, 2025-04-25) Belmonte, José María; Blanquer Blanquer, Miguel; Bernabé García, Gregorio; Jiménez Barrionuevo, Fernando; García Carrasco, José Manuel; Ingeniería y Tecnología de Computadores
This work explores the application of Survival Analysis in the context of hematopoietic stem cell transplantation for multiple myeloma to enhance the predictive capacity and interpretability of transplant outcomes and inspect the patients’ overall survival. Our methodology uses all the proposed Survival Analysis models. These models are used to conduct a feature importance analysis and do some survival predictions with the interpretation of patient outcomes. The dataset, comprising 254 instances and 15 attributes, includes medical information collected from multiple myeloma patients before hematopoietic stem cell transplantation procedures. The primary objective of this work is to assess the robustness of Survival Analysis models with our data based on the concordance index metric. Through feature importance analysis, it has been revealed that variables such as the International Staging System, treatment lines, age, and disease relapse play pivotal roles in determining patient survival post-transplant. Survival predictions have been conducted for three distinct cases from the dataset, evaluating the risks patients may encounter following their treatments. These results have been validated by healthcare professionals, underscoring the reliability and applicability of this study’s findings in medical scenarios.
Open Access
Trabeculated myocardium in hypertrophic cardiomyopathy: clinical consequences
(MDPI, 2020-09-27) Casanova, José David; González Carrillo, Josefa; Martín Jiménez, Jesús; Cuenca Muñoz, Antonio Javier; Muñoz Esparza, Carmen; Siguero Alvárez, Marcos; Escribá, Rubén; Burillo Milla , Esther; Pompa, José Luis de la; Raya, Ángel; Gimeno, Juan Ramón; Bernabé García, Gregorio; Sabater Molina, María; Ingeniería y Tecnología de Computadores
Hypertrophic cardiomyopathy (HCM) is often accompanied by increased trabeculated myocardium (TM)—which clinical relevance is unknown. We aim to measure the left ventricular (LV) mass and proportion of trabeculation in an HCM population and to analyze its clinical implication. Methods and Results: We evaluated 211 patients with HCM (mean age 47.8 ± 16.3 years, 73.0% males) with cardiac magnetic resonance (CMR) studies. LV trabecular and compacted mass were measured using dedicated software for automatic delineation of borders. Mean compacted myocardium (CM) was 160.0 ± 62.0 g and trabecular myocardium (TM) 55.5 ± 18.7 g. The percentage of trabeculated myocardium (TM%) was 26.7% ± 6.4%. Females had significantly increased TM% compared to males (29.7 ± 7.2 vs. 25.6 ± 5.8, p < 0.0001). Patients with LVEF < 50% had significantly higher values of TM% (30.2% ± 6.0% vs. 26.6% ± 6.4%, p = 0.02). Multivariable analysis showed that female gender and neutral pattern of hypertrophy were directly associated with TM%, while dynamic obstruction, maximal wall thickness and LVEF% were inversely associated with TM%. There was no association between TM% with arterial hypertension, physical activity, or symptoms. Atrial fibrillation and severity of hypertrophy were the only variables associated with cardiovascular death. Multivariable analysis failed to demonstrate any correlation between TM% and arrhythmias. Conclusions: Approximately 25% of myocardium appears noncompacted and can automatically be measured in HCM series. Proportion of non-compacted myocardium is increased in female, non-obstructives, and in those with lower contractility. The amount of trabeculation might help to identify HCM patients prone to systolic heart failure.
Open Access
Characterization of machine learning compilers for LLM inference on NVIDIA GPUs
(Springer, 2026-05-15) Bernabé García, Gregorio; José Manuel García; Carmona Martínez, Alejandro; Ingeniería y Tecnología de Computadores; Facultades de la UMU::Facultad de Informática
AI inference is conflicted between Performance, developer Productivity, and device Portability–the P3 problem. Machine learning compilers (MLCs) aim to address this, but their ecosystem is fragmented, with tools that each prioritize a different issue. This paper evaluates the deployment trade-offs of PyTorch-based LLMs on NVIDIA GPUs using four intertwined prominent MLC tools: torch.compile, TensorRT, XLA, and ONNX Runtime. A dual methodology is used, leveraging synthetic PyTorch models to isolate optimizations and end-to-end benchmarks with State-of-the-Art (SOTA) models (TinyLlama-1.1B, Llama-2-7B) to measure realworld performance. Findings reveal that the peak performance of Ahead-Of-Time (AOT) compilation requires architecture-specific tools such as TensorRT-LLM, which are necessary for SOTA LLMs but are unusable for PyTorch models. As for Just-In-Time (JIT) solutions such as torch.compile and its backends, they are flexible and portable, compatible with all tested models, but they do not consistently accelerate LLMs; therefore, the choice of MLC depends on P3 considerations and model architecture.
Open Access
A Data-Centric Algorithmic Pipeline for Enhancing Cardiac MRI Segmentation Using ViTUNeT and Quality-Aware Filtering
(MDPI, 2026-03-06) Salvador de Haro; Jesús Cámara; González Férez, María Pilar; José Manuel García; García Carrasco, José Manuel; Bernabé García, Gregorio; Ingeniería y Tecnología de Computadores; Facultades de la UMU::Facultad de Informática
The performance of deep-learning-based segmentation models is strongly dependent on the quality of the input data, which is frequently heterogeneous or degraded in real-world medical imaging scenarios. This work presents a data-centric algorithmic pipeline designed to improve cardiac MRI segmentation accuracy through systematic image enhancement and automatic slice-quality filtering. The proposed method is formalized as deterministic algorithm that combines image processing and supervised learning components. The approach integrates a contrast- and structure-preserving enhancement stage, based on bilateral filtering and adaptive histogram equalization, with a quality-aware selection algorithm. Slice quality is assessed using anatomical attributes extracted via YOLOv11sbased localization and a supervised classification model trained to identify diagnostically reliable images. When applied to transformer-based segmentation architectures such as ViTUNeT, the pipeline yields consistent improvements across all evaluation metrics without increasing model complexity or training cost. These findings emphasize the importance of algorithmic data curation as an effective strategy for enhancing robustness and stability in deep-learning segmentation pipelines and demonstrate the broader applicability of the proposed approach to computer-vision tasks involving heterogeneous or low-quality image datasets.
Open Access
Sensitivity-constrained evolutionary feature selection for imbalanced medical classification: a case study on rotator cull tear surgery prediction
(MDPI, 2025-12-08) Belmonte, José María; Jiménez Barrionuevo, Fernando; Sánchez Carpena, Gracia; Gabardo, Santiago; Martínez Catalán, Natalia; Calvo, Emilio; Bernabé García, Gregorio; García Carrasco, José Manuel; Ingeniería de la Información y las Comunicaciones; Facultades de la UMU::Facultad de Informática
While most patients with degenerative rotator cuff tears respond to conservative treatment, a minority progress to surgery. To anticipate these cases under class imbalance, we propose a sensitivity-constrained evolutionary feature selection framework prioritizing surgical-class recall, benchmarked against traditional methods. Two variants are proposed: (i) a single-objective search maximizing balanced accuracy and (ii) a multi-objective search also minimizing the number of selected features. Both enforce a minimum-sensitivity constraint on the minority class to limit false negatives. The dataset includes 347 patients (66 surgical, 19%) described by 28 clinical, imaging, symptom, and functional variables. We compare against 62 widely adopted pipelines, including oversampling, undersampling, hybrid resampling, cost-sensitive classifiers, and imbalance-aware ensembles. The main metric is balanced accuracy, with surgical-class F1-score as secondary. PairwiseWilcoxon tests with a win–loss ranking assessed statistical significance. Evolutionary models rank among the top; the multi-objective variant with a Balanced Bagging Classifier performs best, achieving a mean balanced accuracy of 0.741. Selected subsets recurrently include age, tear location/severity, comorbidities, and pain/functional scores, matching clinical expectations. The constraint preserved minority-class recall without discarding or synthesizing data. Sensitivity-constrained evolutionary feature selection thus offers a data-preserving, interpretable solution for pre-surgical decision support, improving balanced performance and supporting safer triage decisions.
Open Access
A Comparative Analysis of Machine Learning and Deep Learning Approaches for Multiclass Nucleus Classifcation in Histological Images
(Wiley, 2026-01-30) Sánchez-Torres, Antonio Luis; García-Salmerón, Jesús ; González Férez, María Pilar; Bernabé García, Gregorio; García Carrasco, José Manuel; Ingeniería y Tecnología de Computadores; Kalapraveen Bagadi; Facultades de la UMU::Facultad de Informática
Precisely classifying cells in histological images is critical for early cancer diagnosis and tumor assessment. Traditional manual methods are time-consuming and labor-intensive for histopathologists, driving the development of automated approaches using machine learning (ML) and deep learning (DL). Convolutional neural networks (CNNs) and, more recently, vision transformers (ViTs) have demonstrated signifcant potential in addressing the challenges of cell classifcation by leveraging their ability to automatically extract and learn complex features from histological images. In this work, we evaluate multiple classifcation architectures applied to stained histological images to determine their efectiveness in identifying cancerous cells. We compare traditional ML models, which rely on manually extracted features such as shape and texture, against two DL-based classifers: a CNN-based model (ResNet50) and a ViT-based model. To optimize ML models, we apply principal component analysis (PCA) to refne feature selection. Meanwhile, DL models are trained on cropped cell images using two preprocessing strategies: one that includes additional surrounding cellular context and another that uses only the cell pixels. Additionally, we investigate class balancing strategies, including downsampling and oversampling through data augmentation, to mitigate the efects of dataset imbalance. Experimental results highlight the clear advantage of DL models over traditional ML approaches. ResNet50 consistently delivers robust and reliable performance across diferent preprocessing strategies, confrming its efectiveness for histopathological classifcation tasks. Meanwhile, ViTs achieve results that are comparable to those of CNNs while demonstrating a distinct advantage in classifying underrepresented nucleus classes, likely due to their ability to capture long-range dependencies. Furthermore, incorporating the surrounding cellular environment signifcantly improves classifcation accuracy, underscoring the importance of contextual information in distinguishing between diferent types of nuclei.