Myocardium segmentation is critical for remote diagnosis and intelligent monitoring of cardiac disease in e-Health systems. Existing deep-learning methods struggle with class imbalance, anatomical heterogeneity, and blurred boundaries in medical images. Standard cross-entropy loss weight all pixels equally and cannot adapt to rare but important regions. Focal loss tend to overly emphasize difficult-to-classify samples and lose edge information. A shape-aware adaptive weighted cross-entropy (AWCE) loss is introduced, which combines a global shape prior with per-pixel uncertainty to focus learning on rare in the dataset but critical regions for each samples. On the ACDC dataset, the proposed AWCE loss achieves stable training convergence, recovers thin-wall structures in qualitative evaluations, and outperforms both cross-entropy and focal loss on key metrics. This shape-guided loss promises to accelerate the adoption of AI-driven cardiac analysis in future digital health applications.

Multi-model fitting is a fundamental challenge in computer vision, where real-world data often contains severe gross outliers and pseudo-outliers. Existing methods rely on inefficient sequential hypothesize-and-verify frameworks that require a predefined number of models and inlier thresholds-parameters that are difficult to determine in practical scenes. To overcome these limitations, we propose a novel Adaptive Graph Attention-guided parallel multi-model fitting method (AGASAC) that jointly learns local and global features, performs parallel hypothesis sampling, and executes confidence-embedded model selection. Specifically, we design a dual-confidence graph attention module that models data relationships using an adaptive graph attention network. This module computes minimal-set confidence and quality confidence to guide the multi-model fitting process, eliminating manual parameter tuning. Additionally, we propose a parallel discriminative sampling module that leverages minimal-set confidence to concurrently sample hypotheses. By enforcing a quantized consensus constraint, this module maximizes inter-model variance while minimizing intra-model discrepancy. It enables computationally efficient hypothesis generation and pseudo-outlier suppression. To obtain high-quality models, we present a quality-embedded selection module that integrates quality confidence into the joint optimization of model selection and data clustering. Extensive experiments show that the proposed method achieves a lower transfer error of 0.39 pixels and a 36.92% runtime reduction, surpassing state-of-the-art methods. The code is available at https://github.com/YWY-Vivian/AGASAC.

Accurate segmentation of the left ventricular myocardium in cardiac CT images is essential for diagnosing and treating cardiovascular diseases. This is considered particularly critical in clinical workflows enabled by the Internet of Medical Things (IoMT), where real-time remote monitoring and rapid analysis of cardiac function are fundamental to timely and effective intervention. Deep learning techniques have achieved significant progress in automatically outlining target objects with reduced human effort. However, most of the deep learning approaches focus on pixel-level loss minimization in model training, and fail to capture the global shape characteristics of anatomical structures, resulting in inconsistent segmentation. In this paper, we propose an end-to-end deep learning framework, which incorporates shape priors derived using kernel principal component analysis (KPCA) to address such limitation. A shape loss function is introduced to enforce anatomical shape consistency by integrating shape prior of myocardial structures during model training. Comprehensive experiments on both convolutional and transformer-based architectures demonstrate that the proposed method significantly enhances segmentation accuracy and anatomical fidelity, particularly in regions critical for diagnosis, such as mitral valve, interventricular septum and papillary muscles. Moreover, we show that integration of learnable Gaussian kernel parameters in KPCA enhances adaptability, outperforming that with fixed kernel parameters. We believe that the improved segmentation accuracy helps doctors make more precise assessments of the heart. This is especially useful in IoMT-based healthcare systems, where reliable segmentation supports quick diagnosis, remote monitoring, and smarter medical decisions, leading to better care and outcomes for patients with heart disease.

Multi-model fitting aims to robustly estimate the parameters of various model instances in data contaminated by noise and outliers. Most previous works employ only a single type of consensus or implicit fusion model to represent the correlation between data points and model hypotheses. This approach often results in unrealistic and incorrect model fitting in the presence of noise and uncertainty. In this paper, we propose a novel method of diverse Consensuses paired with Motion estimation-based multi-Model Fitting (CMMF), which leverages three types of diverse consensuses along with inter-model collaboration to enhance the effectiveness of multi-model fusion. We design a Tangent Consensus Residual Reconstruction (TCRR) module to capture motion structure information of two points at the pixel level. Additionally, we introduce a Cross Consensus Affinity (CCA) framework to strengthen the correlation between data points and model hypotheses. To address the challenge of multi-body motion estimation, we propose a Nested Consensus Clustering (NCC) strategy, which formulates multi-model fitting as a motion estimation problem. It explicitly establishes motion collaboration between models and ensures that multiple models are well-fitted. Extensive quantitative and qualitative experiments are conducted on four public datasets (i.e., AdelaideRMF-F, Hopkins155, KITTI, MTPV62), and the results demonstrate that our proposed method outperforms several state-of-the-art methods.