Cardiovascular Disease (CVD) is the most common cause of mortality worldwide, so reliable tools are needed to accurately diagnose it to provide timely clinical interventions. Traditional forms of diagnosis relied solely on single-modality data, used basic fusion approaches at diagnosis, or did not consider how to complement data across heterogeneous modalities. This study presents a novel Hierarchical Attention-based Representation learning with Multi-modal Network (HARM-Net) framework to grade the severity of CVD. The proposed framework combines medical imaging, physiological signals, electronic health records, and demographic data. It consists of a five-step process that includes using modality-specific encoders, self-supervised pre-training via multi-modal contrastive learning, hierarchical cross-attention fusion, compressing deep features, and a modality-specific adaptive stacking ensemble classification model. Extensive experimental results on the MultiD4CAD dataset demonstrated that the HARM-Net framework outperformed other models with an accuracy of 93.2%, an F1-score of 91.7%, and a ROC-AUC of 0.967. The results of an ablation study demonstrate the significance of each component and the specific contribution of cross-attention fusion to improved model performance for accurate, multimodal diagnosis of CVD.
S. Shwetha, Y. Manu· Engineering, Technology &...· 0 citations
Segmentation of brain tumors in Magnetic Resonance Imaging (MRI) has many applications in the diagnosis, treatment planning, and monitoring of the disease. The manual delineation of tumor sub-regions is time-consuming and can be influenced by inter-observer variability, requiring automated solutions. This study introduces an AI-based 3D brain tumor segmentation system based on a 3D U-Net architecture to segment brain tumors based on multi-classes with multi-modal MRI volumetric data. Four MRI modalities are processed, and the tumor regions are divided into edema, necrotic core, and enhancing tumor in the proposed model. To address class imbalance and enhance segmentation accuracy, a loss function based on the Dice coefficient is used for training. Besides segmentation, the system consists of quantitative tissue analysis via calculation of the relative distribution of tumor sub-regions and interactive visualization using a web-based dashboard. The experiments demonstrated successful volumetric segmentation and significant tissue quantification. The proposed framework emphasizes a system-level integration of segmentation, quantification, and visualization to enhance interpretability and practical usability.
D. U. Latha, M. Padma, D. Rajeshwari et al.· Engineering, Technology &...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
