Cai Meng

Segmentation of the pancreatic computed tomography (CT) image in the Internet of Medical Things (IoMT) environment faces dual challenges of feature robustness and accurate recognition of small organs. As a clinically critical but anatomically challenging organ, the pancreas exhibits small volume, high interpatient variability, and low contrast with surrounding tissues, making automatic pancreas segmentation a representative and difficult task in abdominal CT analysis. Existing automatic, machine-centric segmentation methods often perform unsatisfactorily across different medical institutions due to variations in imaging devices, changes in scanning protocols, as well as the irregular shape and blurred boundaries of the pancreas. To address this problem, this article proposes a SimCLR-based Contrastive Swin-UNet (ContSwinU) model, which integrates contrastive learning with the Swin Transformer architecture to achieve feature robustness learning and high-precision pancreas segmentation. Specifically, ContSwinU leverages SimCLR to learn robust feature representations, thereby enhancing the model’s generalization capability across diverse scenarios. Additionally, by incorporating the hierarchical window attention mechanism of the Swin Transformer, the model effectively balances local texture and global structural information, improving segmentation accuracy for the pancreas as a small organ. Experimental results on a public pancreas CT dataset demonstrate that ContSwinU achieves an IoU of 0.8041, a Dice coefficient of 0.8872, and a recall of 0.9014, significantly outperforming mainstream baseline methods. These results indicate that the proposed framework is well suited for challenging small-organ segmentation tasks and has the potential to be extended to other organs and multicenter IoMT scenarios. This study provides an effective solution for pancreas segmentation in IoMT environments and has substantial clinical application value.

Pancreatic cancer is one of the most aggressive malignant solid tumors, and achieving early screening and diagnosis is the key to improving patient survival rates. Although deep learning has made significant progress in medical image analysis, the classification of pancreatic cancer computed tomography (CT) images remains highly challenging due to subtle interlesion differences and overlapping category distributions. Moreover, existing methods rely heavily on large amounts of high-quality annotations, which can lead to overfitting and hinder the effective exploitation of structural and semantic features within images. To address these challenges, we propose a contrastive Swin transformer with masked autoencoder (CSTMA) for pancreatic cancer CT image classification in the Internet of Medical Things (IoMT) environment. CSTMA leverages contrastive learning to enhance feature discriminability, while its multitask self-supervised architecture based on masked autoencoder (MAE) guides the model to learn both structural and semantic representations. We conduct comprehensive experiments on pancreatic cancer CT image classification tasks, and the results demonstrate that the proposed CSTMA model achieves superior performance across multiple evaluation metrics.

Accurate vessel segmentation is crucial for diagnosing cerebrovascular and retinal diseases, but the complexity of vessel structures makes manual annotation challenging, resulting in a lack of annotated data. Additionally, generated data often struggle to balance realism and diversity in vessel morphology. To address this, we provide a theoretical description of vessel structure and propose a vascular morphology-driven augmentation pipeline (VMDAP) for both feature reconstruction and vessel generation based on real vascular morphological characteristics. The pipeline consists of three components: the background restoration block (BRB), the vessel structure feature enhancement block (VSFEB), and the augmented data generation network (ADGN). VMDAP strategically separates vessel data into foreground and background, restores the background, reconstructs vessel masks features, and reintegrates them to generate new data. The ADGN incorporates vessel texture priors and employs a vessel detail enhancement loss to improve the realism of details. To address the performance characteristics of the VMDAP algorithm, this paper also introduces a novel loss function, DAclDice, which focuses on vessel branch structures to reduce false positives. Extensive experiments on six datasets across three representative network architectures validate the effectiveness of VMDAP and structure-driven loss function (DAclDice). Compared to nnU-Net with clDice loss, our method achieves an average Dice coefficient improvement of 2.15, a clDice improvement of 1.16, and a significant 36.17% reduction in 95th percentile Hausdorff distance (HD95). Excitingly, additional experiments demonstrate that VMDAP outperforms full real-data training when only half the real data are used.

Accurate 6D pose estimation of 3D objects is a fundamental task in computer vision, and current research typically predicts the 6D pose by establishing correspondences between 2D image features and 3D model features. However, these methods often face difficulties with textureless objects and varying illumination conditions. To overcome these limitations, we propose GS2POSE, a novel approach for 6D object pose estimation. GS2POSE formulates a pose regression algorithm inspired by the principles of Bundle Adjustment (BA). By leveraging Lie algebra, we extend the capabilities of 3DGS to develop a pose-differentiable rendering pipeline, which iteratively optimizes the pose by comparing the input image to the rendered image. Additionally, GS2POSE updates color parameters within the 3DGS model, enhancing its adaptability to changes in illumination. Compared to previous models, GS2POSE demonstrates accuracy improvements of 1.4\%, 2.8\% and 2.5\% on the T-LESS, LineMod-Occlusion and LineMod datasets, respectively.

Precise segmentation of thermal infrared images is crucial in domains like surveillance, medical diagnostics, intelligent transportation, accurate guidance and remote sensing. However, current thermal segmentation methods often oversimplify by treating thermal images as grayscale, neglecting vital physical factors such as thermal imaging effects and material information, thereby constraining segmentation precision. To address these limitations, we propose TherNet, a novel thermal infrared segmentation framework integrating thermal imaging effects and material physical information. The study elucidates the impacts of object radiation, inter-object thermal exchange, atmospheric scattering, and camera thermal inertia on thermal infrared imaging, developing four modules to model or rectify these physical processes. To validate the proposed framework, two large-scale infrared datasets were created: TI-Cityscapes for multi-class semantic segmentation in traffic scenes (4,200 frames, 18 classes), and TBRSD for single-object blindroad segmentation (5,180 frames from a pedestrian perspective). The proposed methods achieved SoTA performance across three infrared semantic segmentation datasets and the blind road segmentation dataset, underscoring the pivotal role of leveraging physical properties. TherNet provides innovative perspectives and robust benchmarks for future developments in the domain.