International Journal of Imaging Systems and Technology最新文献

Automated diagnosis of retinal diseases holds significant promise in enhancing healthcare efficiency and patient outcomes. However, existing methods often lack the accuracy and efficiency required for timely disease detection. To address this gap, we introduce RetNet30, a novel stacked convolutional neural network (CNN) designed to revolutionize automated retinal disease diagnosis. RetNet30 combines a custom-built 30-layer CNN with a fine-tuned Inception V3 model, integrating these sub-models through logistic regression to achieve superior classification performance. Extensive evaluations on retinal image datasets such as DRIVE, STARE, CHASE_DB1, and HRF demonstrate significant improvements in accuracy, sensitivity, specificity, and area under the ROC curve (AUROC) when compared to conventional approaches. By leveraging advanced deep learning architectures, RetNet30 not only enhances diagnostic precision but also generalizes effectively across diverse datasets, establishing a new benchmark in retinal disease classification. This novel approach offers a highly efficient and reliable solution for early disease detection and patient management, addressing the limitations of manual examination methods. Through rigorous quantitative and qualitative assessments, our proposed method demonstrates its potential to significantly impact medical image analysis and improve healthcare outcomes. RetNet30 marks a major step forward in automated retinal disease diagnosis, showcasing the future of AI-driven advancements in ophthalmology.

Accurately and rapidly segmenting the prostate in transrectal ultrasound (TRUS) images remains challenging due to the complex semantic information in ultrasound images. The paper discusses a cross-layer connection with SegFormer attention U-Net for efficient TRUS image segmentation. The SegFormer framework is enhanced by reducing model parameters and complexity without sacrificing accuracy. We introduce layer-skipping connections for precise positioning and combine local context with global dependency for superior feature recognition. The decoder is improved with Multi-layer Perceptual Convolutional Block Attention Module (MCBAM) for better upsampling and reduced information loss, leading to increased accuracy. The experimental results show that compared with classic or popular deep learning methods, this method has better segmentation performance, with the dice similarity coefficient (DSC) of 97.55% and the intersection over union (IoU) of 95.23%. This approach balances encoder efficiency, multi-layer information flow, and parameter reduction.

Colorectal adenocarcinoma, the most prevalent form of colon cancer, originates in the glandular structures of the intestines, presenting histopathological abnormalities in affected tissues. Accurate gland segmentation is crucial for identifying these potentially fatal abnormalities. While recent methodologies have shown success in segmenting glands in benign tissues, their efficacy diminishes when applied to malignant tissue segmentation. This study aims to develop a robust learning algorithm using a convolutional neural network (CNN) to segment glandular structures in colon histology images. The methodology employs a CNN based on the U-Net architecture, augmented by a weighted ensemble network that integrates DenseNet 169, Inception V3, and Efficientnet B3 as backbone models. Additionally, the segmented gland boundaries are refined using the watershed algorithm. Evaluation on the Warwick-QU dataset demonstrates promising results for the ensemble model, by achieving an F1 score of 0.928 and 0.913, object dice coefficient of 0.923 and 0.911, and Hausdorff distances of 38.97 and 33.76 on test sets A and B, respectively. These results are compared with outcomes from the GlaS challenge (MICCAI 2015) and existing research findings. Furthermore, our model is validated with a publicly available dataset named LC25000, and visual inspection reveals promising results, further validating the efficacy of our approach. The proposed ensemble methodology underscores the advantages of amalgamating diverse models, highlighting the potential of ensemble techniques to enhance segmentation tasks beyond individual model capabilities.

Diagnosing diffuse infiltrative lung diseases (DILD) using high-resolution computed tomography (HRCT) is challenging, even for expert radiologists, due to the complex and variable image patterns. Moreover, the imbalances among the six key DILD-related patterns—normal, ground-glass opacity, reticular opacity, honeycombing, emphysema, and consolidation—further complicate accurate segmentation and diagnosis. This study presents an enhanced U-Net-based segmentation technique aimed at addressing these challenges. The primary contribution of our work is the fine-tuning of the U-Net model using image-level labels from 92 HRCT images that include various types of DILDs, such as cryptogenic organizing pneumonia, usual interstitial pneumonia, and nonspecific interstitial pneumonia. This approach helps to correct the imbalance among image patterns, improving the model's ability to accurately differentiate between them. By employing semantic lung segmentation and patch-level machine learning, the fine-tuned model demonstrated improved agreement with radiologists' evaluations compared to conventional methods. This suggests a significant enhancement in both segmentation accuracy and inter-observer consistency. In conclusion, the fine-tuned U-Net model offers a more reliable tool for HRCT image segmentation, making it a valuable imaging biomarker for guiding treatment decisions in patients with DILD. By addressing the issue of pattern imbalances, our model significantly improves the accuracy of DILD diagnosis, which is crucial for effective patient care.

In the clinical application of the parallel external fixator, medical practitioners are required to quantify deformity parameters to develop corrective strategies. However, manual measurement of deformity angles is a complex and time-consuming process that is susceptible to subjective factors, resulting in nonreproducible results. Accordingly, this study proposes an automatic measurement method based on deep learning, comprising three stages: tibial segment localization, tibial contour point detection, and deformity angle calculation. First, the Faster R-CNN object detection model, combined with ResNet50 and FPN as the backbone, was employed to achieve accurate localization of tibial segments under both occluded and nonoccluded conditions. Subsequently, a relative position constraint loss function was added, and ResNet101 was used as the backbone, resulting in an improved RTMPose keypoint detection model that achieved precise detection of tibial contour points. Ultimately, the bone axes of each tibial segment were determined based on the coordinates of the contour points, and the deformity angles were calculated. The enhanced keypoint detection model, Con_RTMPose, elevated the Percentage of Correct Keypoints (PCK) from 63.94% of the initial model to 87.17%, markedly augmenting keypoint localization precision. Compared to manual measurements conducted by medical professionals, the proposed methodology demonstrates an average error of 0.52°, a maximum error of 1.15°, and a standard deviation of 0.07, thereby satisfying the requisite accuracy standards for orthopedic assessments. The measurement time is approximately 12 s, whereas manual measurement requires about 15 min, greatly reducing the time required. Additionally, the stability of the models was verified through K-fold cross-validation experiments. The proposed method meets the accuracy requirements for orthopedic applications, provides objective and reproducible results, significantly reduces the workload of medical professionals, and greatly improves efficiency.

The problem of missing or unavailable magnetic resonance imaging modalities challenges clinical diagnosis and medical image analysis technology. Although the development of deep learning and the proposal of large models have improved medical analytics, this problem still needs to be better resolved.The purpose of this study was to efficiently adapt the Segment Anything Model, a two-dimensional visual foundation model trained on natural images, to address the challenge of brain tumor segmentation with missing modalities. We designed a twin network structure that processes missing and intact magnetic resonance imaging (MRI) modalities separately using shared parameters. It involved comparing the features of two network branches to minimize differences between the feature maps derived from them. We added a multimodal adapter before the image encoder and a spatial–depth adapter before the mask decoder to fine-tune the Segment Anything Model for brain tumor segmentation. The proposed method was evaluated using datasets provided by the MICCAI BraTS2021 Challenge. In terms of accuracy and robustness, the proposed method is better than existing solutions. The proposed method can segment brain tumors well under the missing modality condition.

Accurate segmentation of brain tumors from magnetic resonance imaging (MRI) is crucial for diagnosis, treatment planning, and monitoring of patients with neurological disorders. This paper proposes an approach for brain tumor segmentation employing a cascaded architecture integrating L-Net and W-Net deep learning models. The proposed cascaded model leverages the strengths of U-Net as a baseline model to enhance the precision and robustness of the segmentation process. In the proposed framework, the L-Net excels in capturing the mask, while the W-Net focuses on fine-grained features and spatial information to discern complex tumor boundaries. The cascaded configuration allows for a seamless integration of these complementary models, enhancing the overall segmentation performance. To evaluate the proposed approach, extensive experiments were conducted on the datasets of BraTs and SMS Medical College comprising multi-modal MRI images. The experimental results demonstrate that the cascaded L-Net and W-Net model consistently outperforms individual models and other state-of-the-art segmentation methods. The performance metrics such as the Dice Similarity Coefficient value achieved indicate high segmentation accuracy, while Sensitivity and Specificity metrics showcase the model's ability to correctly identify tumor regions and exclude healthy tissues. Moreover, the low Hausdorff Distance values confirm the model's capability to accurately delineate tumor boundaries. In comparison with the existing methods, the proposed cascaded scheme leverages the strengths of each network, leading to superior performance compared to existing works of literature.