Electrocardiogram (ECG) interpretation using deep learning models holds immense potential for improving cardiac diagnosis. However, existing models often suffer from overconfident predictions and lack the capability to directly quantify uncertainty, leading to unreliable clinical guidance. To address this challenge, we propose a model for uncertainty-aware ECG interpretation. The model employs a deep convolutional architecture with max-pooling residual modules to capture both local and global spatiotemporal features from raw ECG signals. The architectural design respects the symmetry inherent in ECG waveforms—such as periodicity and morphological consistency across cardiac cycles—enabling the network to extract clinically relevant features more effectively. Then, unlike conventional models that rely on softmax-based probability outputs, our approach parameterizes class distributions using the Dirichlet distribution, while Subjective Logic translates these parameters into interpretable belief masses and uncertainty scores. We evaluate the model on the PhysioNet Challenge 2017 dataset, our model achieves an accuracy of 86.12%, an F1 score of 83.14%, a Precision-Recall Area Under the Curve (PR-AUC) of 85.25%, and a Receiver Operating Characteristic Area Under the Curve (ROC-AUC) of 92.87%—outperforming baseline models in three out of four metrics. Critically, the model reduces overconfidence to 0.59% (compared to 12–22% in softmax-based baselines), aligning prediction confidence with true accuracy. By progressively increasing the uncertainty threshold u, the model dynamically filters low-confidence predictions, leading to consistently improved performance—reaching up to 93.59% accuracy, 93.22% F1 score, 89.17% PR-AUC, and 95.10% ROC-AUC at u = 0.1. These results validate the model’s capacity for reliable ECG interpretation while leveraging physiological signal symmetry for enhanced feature extraction.

Pressure injury management requires coordinated decisions on wound staging, debridement, and dressing selection, yet most computational studies focus on isolated tasks using a single data modality. This study developed a multimodal, multi-task framework integrating wound photographs and clinical text to support pressure injury stage classification, debridement necessity prediction, primary dressing recommendation, and secondary dressing recommendation. Models were trained on an expert-annotated public dataset and externally validated using an independent private cohort. We systematically compared image-only models, text-only models, large language model-based approaches, conventional image–text fusion strategies, and image–language model combinations. Among the evaluated models, ResNet-50 combined with DeepSeek-R1-Distill-Qwen-1.5B showed the best overall balance across classification performance, text-generation quality, and validation robustness. On the validation set, the final model achieved 82.92% accuracy, 81.26% macro-F1, and 99.01% area under the curve for stage classification, and 89.91% accuracy, 89.83% macro-F1, and 96.24% area under the curve for debridement prediction. For dressing recommendation, the model achieved ROUGE-1 scores of 58.30% and 61.06% for primary and secondary dressings, respectively. Interpretability analyses indicated clinically relevant image attention and text attribution patterns. These findings suggest that multimodal learning may provide a clinically aligned decision-support approach for pressure injury assessment and treatment recommendations.