ct体数据异常检测及ct重建联合检测

Automated pulmonary anomaly detection using computed tomography (CT) examinations is important for the early warning of pulmonary diseases and can support clinical diagnosis and decision-making. Most training of existing pulmonary disease detection and lesion segmentation models requires expert annotations, which is time-consuming and labour-intensive, and struggles to generalize across atypical diseases. In contrast, unsupervised anomaly detection alleviates the demand for dataset annotation and is more generalizable than supervised methods in detecting rare pathologies. However, due to the large distribution differences of CT scans in a volume and the high similarity between lesion and normal tissues, existing anomaly detection methods struggle to accurately localize small lesions, leading to a low anomaly detection rate. To alleviate these challenges, we propose a local salient location-aware anomaly mask generation and reconstruction framework for pulmonary disease anomaly detection and lesion localization. The framework consists of four components: (1) a Vector Quantized Variational AutoEncoder (VQVAE)-based reconstruction network that generates a codebook storing high-dimensional features; (2) a unsupervised feature statistics based anomaly feature synthesizer to synthesize features that match the realistic anomaly distribution by filtering salient features and interacting with the codebook; (3) a transformer-based feature classification network that identifies synthetic anomaly features; (4) a residual neighbourhood aggregation feature classification loss that mitigates network overfitting by penalizing the classification loss of recoverable corrupted features. Our approach is based on two intuitions. First, generating synthetic anomalies in feature space is more effective due to the fact that lesions have different morphologies in image space and may not have much in common. Secondly, regions with salient features or high reconstruction errors in CT images tend to be similar to lesions and are more prone to synthesize abnormal features. The performance of the proposed method is validated on one public dataset with COVID-19 and one in-house dataset containing 63,610 CT images with five lung diseases. Experimental results show that compared to feature-based, synthesis-based and reconstruction-based methods, the proposed method is adaptable to CT images with four pneumonia types (COVID-19, bacteria, fungal, and mycoplasma) and one non-pneumonia (cancer) diseases and achieves state-of-the-art performance in image-level anomaly detection and lesion localization.

In this work, we present a novel latent diffusion-based pipeline for 3D kidney anomaly detection on contrast-enhanced abdominal CT. The method combines Denoising Diffusion Probabilistic Models (DDPMs), Denoising Diffusion Implicit Models (DDIMs), and Vector-Quantized Generative Adversarial Networks (VQ-GANs). Unlike prior slice-wise approaches, our method operates directly on an image volume and leverages weak supervision with only case-level pseudo-labels. We benchmark our approach against state-of-the-art supervised segmentation and detection models. This study demonstrates the feasibility and promise of 3D latent diffusion for weakly supervised anomaly detection. While the current results do not yet match supervised baselines, they reveal key directions for improving reconstruction fidelity and lesion localization. Our findings provide an important step toward annotation-efficient, generative modeling of complex abdominal anatomy.

Abstract. Purpose To help radiologists examine the growing number of computed tomography (CT) scans, automatic anomaly detection is an ongoing focus of medical imaging research. Radiologists must analyze a CT scan by searching for any deviation from normal healthy anatomy. We propose an approach to detecting abnormalities in axial 2D CT slice images of the brain. Although much research has been done on detecting abnormalities in magnetic resonance images of the brain, there is little work on CT scans, where abnormalities are more difficult to detect due to the low image contrast that must be represented by the model used. Approach We use a generative adversarial network (GAN) to learn normal brain anatomy in the first step and compare two approaches to image reconstruction: training an encoder in the second step and using iterative optimization during inference. Then, we analyze the differences from the original scan to detect and localize anomalies in the brain. Results Our approach can reconstruct healthy anatomy with good image contrast for brain CT scans. We obtain median Dice scores of 0.71 on our hemorrhage test data and 0.43 on our test set with additional tumor images from publicly available data sources. We also compare our models to a state-of-the-art autoencoder and a diffusion model and obtain qualitatively more accurate reconstructions. Conclusions Without defining anomalies during training, a GAN-based network was used to learn healthy anatomy for brain CT scans. Notably, our approach is not limited to the localization of hemorrhages and tumors and could thus be used to detect structural anatomical changes and other lesions.

Coronary artery disease is a leading cause of mortality world-wide, emphasizing the need for accurate and efficient detection methods. Current deep learning approaches for coronary lesion detection in computed tomography coronary angiog-raphy (CTCA) rely on large labeled datasets for supervised learning, limiting scalability. This study proposes an unsu-pervised framework for coronary lesion detection in CTCA. Using vessel linearization, we generated small linear volume patches from CTCA and trained a deep autoencoder model, MemAE, exclusively on normal patches to detect anomalies based on reconstruction error without lesion location labels. Results on a public dataset demonstrate high accuracy, achieving 97.5% on per-patient average reconstruction error and 85% using the number of patches with high reconstruction error, offering a scalable and accessible solution.

Anomaly detection has been widely explored by training an out-of-distribution detector with only normal data for medical images. However, detecting local and subtle irregularities without prior knowledge of anomaly types brings challenges for lung CT-scan image anomaly detection. In this paper, we propose a self-supervised framework for learning representations of lung CT-scan images via both multi-scale cropping and simple masked attentive predicting, which is capable of constructing a powerful out-of-distribution detector. Firstly, we propose CropMixPaste, a self-supervised augmentation task for generating density shadow-like anomalies that encourage the model to detect local irregularities of lung CT-scan images. Then, we propose a self-supervised reconstruction block, named simple masked attentive predicting block (SMAPB), to better refine local features by predicting masked context information. Finally, the learned representations by self-supervised tasks are used to build an out-of-distribution detector. The results on real lung CT-scan datasets demonstrate the effectiveness and superiority of our proposed method compared with state-of-the-art methods.

Detecting subtle focal liver lesions on abdominal computed tomography (CT) is challenging in routine clinical practice, especially for small, low-contrast, or morphologically heterogeneous tumors acquired under variable protocols. While fully supervised liver tumor segmentation can achieve high accuracy, it requires pixel-level annotations that limit scalability and generalizability. Reconstruction-based anomaly detectors trained without hepatic anatomical constraints reduce label burden but are sensitive to textural variability, contrast-phase differences, and produce noisy, unstable boundaries. We introduce an anatomically constrained, four-stage pipeline for liver CT anomaly detection: (1) a denoising diffusion probabilistic model (DDPM) trained on unremarkable axial slices to learn a healthy prior; (2) diffusion-based inpainting within an automatically segmented whole-liver mask to generate pseudo-normal liver appearance; (3) a compact encoder–decoder trained with a liver-masked, mean squared error loss to reconstruct healthy liver tissue from paired original and inpainted inputs; and (4) a liver-scoped difference map between the original and reconstructed healthy CT slices as the final anomaly score for localization. Trained exclusively on > 13,000 healthy CT slices and evaluated on 1,000 abnormal CT slices from 109 Liver Tumor Segmentation (LiTS) benchmark patients, the method achieves Dice 0.596, intersection-over-union 0.482, area under the receiver operating characteristic curve 0.861, and 95th percentile Hausdorff distance 80.5 pixels (px). Performance improves with lesion size, with a Dice score of 0.796 for the largest quartile. Anchoring anomaly detection to hepatic anatomy with a stable healthy prior yields data-efficient liver lesion localization suitable for CT triage and prioritization.

Background The early identification of abnormal pulmonary infectious diseases (APIDs) could effectively control the large-scale spread of such diseases. This study proposed a deep learning-based method for the early identification of APIDs. Methods Unsupervised anomaly detection (UAD) refers to the identification of abnormal samples of which its distribution differs from that of normal samples using a training set comprised of only normal samples. Building on this principle, we proposed a method for the early identification of APIDs. First, we established a pulmonary infection computed tomography (PICT) image sequence dataset, which included computed tomography (CT) image sequences of various common pulmonary infections, as well as two known abnormal cases [coronavirus disease 2019 (COVID-19) and melioidosis pneumonia]. Under our framework, only common infection sequences were used to train the UAD network, while both common and abnormal sequences were used in testing to assess the capability of the network to identify deviations. This approach not only detected the two known abnormal cases but was also able to detect unknown APIDs. To enhance the detection accuracy (ACC) of our approach, we developed the local reconstruction autoencoder (LRAE), which focuses on local regions in PICT images to effectively distinguish between common and abnormal infection areas. Results Comprehensive experiments on the PICT dataset were conducted using metrics such as the area under the curve (AUC), F1-score, and ACC, and the results revealed the effectiveness and superiority of the LRAE compared to existing UAD methods. Specifically, the AUC, F1-score, and ACC of the LRAE in detecting COVID-19 CT image sequences were 0.8269, 0.7242, and 0.7801, respectively; while those for the melioidosis pneumonia CT image sequences were 0.8716, 0.6415, and 0.8146, respectively. Conclusions This work offers a robust solution for the early identification of both known and emerging APIDs. The developed LRAE showed remarkable performance in detecting abnormal PICT image sequences.

Anomaly detection in medical imaging plays a crucial role in identifying pathological regions across various imaging modalities, such as brain MRI, liver CT, and carotid ultrasound (US). However, training fully supervised segmentation models is often hindered by the scarcity of expert annotations and the complexity of diverse anatomical structures. To address these issues, we propose a novel unsupervised anomaly detection framework based on a diffusion model that incorporates a synthetic anomaly (Synomaly) noise function and a multi-stage diffusion process. Synomaly noise introduces synthetic anomalies into healthy images during training, allowing the model to effectively learn anomaly removal. The multi-stage diffusion process is introduced to progressively denoise images, preserving fine details while improving the quality of anomaly-free reconstructions. The generated high-fidelity counterfactual healthy images can further enhance the interpretability of the segmentation models, as well as provide a reliable baseline for evaluating the extent of anomalies and supporting clinical decision-making. Notably, the unsupervised anomaly detection model is trained purely on healthy images, eliminating the need for anomalous training samples and pixel-level annotations. We validate the proposed approach on brain MRI, liver CT datasets, and carotid US. The experimental results demonstrate that the proposed framework outperforms existing state-of-the-art unsupervised anomaly detection methods, achieving performance comparable to fully supervised segmentation models in the US dataset. Ablation studies further highlight the contributions of Synomaly noise and the multi-stage diffusion process in improving anomaly segmentation. These findings underscore the potential of our approach as a robust and annotation-efficient alternative for medical anomaly detection. Code:https://github.com/yuan-12138/Synomaly.

To compare the image quality and pulmonary nodule detectability and measurement accuracy between deep learning reconstruction (DLR) and hybrid iterative reconstruction (HIR) of chest ultra-low-dose CT (ULDCT). Participants who underwent chest standard-dose CT (SDCT) followed by ULDCT from October 2020 to January 2022 were prospectively included. ULDCT images reconstructed with HIR and DLR were compared with SDCT images to evaluate image quality, nodule detection rate, and measurement accuracy using a commercially available deep learning–based nodule evaluation system. Wilcoxon signed-rank test was used to evaluate the percentage errors of nodule size and nodule volume between HIR and DLR images. Eighty-four participants (54 ± 13 years; 26 men) were finally enrolled. The effective radiation doses of ULDCT and SDCT were 0.16 ± 0.02 mSv and 1.77 ± 0.67 mSv, respectively (P < 0.001). The mean ± standard deviation of the lung tissue noises was 61.4 ± 3.0 HU for SDCT, 61.5 ± 2.8 HU and 55.1 ± 3.4 HU for ULDCT reconstructed with HIR-Strong setting (HIR-Str) and DLR-Strong setting (DLR-Str), respectively (P < 0.001). A total of 535 nodules were detected. The nodule detection rates of ULDCT HIR-Str and ULDCT DLR-Str were 74.0% and 83.4%, respectively (P < 0.001). The absolute percentage error in nodule volume from that of SDCT was 19.5% in ULDCT HIR-Str versus 17.9% in ULDCT DLR-Str (P < 0.001). Compared with HIR, DLR reduced image noise, increased nodule detection rate, and improved measurement accuracy of nodule volume at chest ULDCT. Not applicable.

This study aimed to evaluate the diagnostic performance, image quality, and radiation dose among ultralow-dose protocol with deep learning reconstruction (DLR), ultralow-dose computed tomography (CT) with iterative reconstruction (IR), and conventional-dose protocols for detecting intracranial hemorrhage. This retrospective study enrolled 93 patients (median age: 67 years; interquartile range [IQR]: 59–76 years; 61 males). A conventional-dose CT was obtained using 120 kVp, 123–188 mA and IR. Follow-up ultralow-dose CT was obtained using 120 kVp, 50 mA with IR and DLR. Qualitative assessments and quantitative assessments were conducted. The diagnostic performance for detecting intracranial hemorrhage was assessed. An approximately 84.0% reduction in median volume CT dose index was found in the ultralow-dose CT protocol (5.6 mGy) compared with conventional-dose CT (35.02 mGy). Ultralow-dose CT with DLR significantly (p < 0.001) reduced image noise, improved signal-to-nosie ratio, and contrast-to-tnoise ratio compared with ultralow-dose CT with IR and conventional-dose CT. Ultralow-dose CT with DLR resulted in higher sensitivity (99.3% vs. 98.6%) and specificity (97.5% vs. 97.5%) for detecting intracranial hemorrhage than ultralow-dose CT with IR. Ultralow-dose CT with DLR is not inferior to conventional-dose CT in terms of image quality and diagnostic performance for the detection of intracranial hemorrhage, while achieving an approximate 87.7% reduction in radiation dose.

No abstract available

The aim of this study is to evaluate cosmetic outcomes of the reconstructed breast in breast cancer patients, using anomaly score (AS) detected by generative adversarial network (GAN) deep learning algorithm. A total of 251 normal breast images from patients who underwent breast-conserving surgery were used for training anomaly GAN network. GAN-based anomaly detection was used to calculate abnormalities as an AS, followed by standardization by using z-score. Then, we reviewed 61 breast cancer patients who underwent mastectomy followed by reconstruction with autologous tissue or tissue expander. All patients were treated with adjuvant radiation therapy (RT) after reconstruction and computed tomography (CT) was performed at three-time points with a regular follow-up; before RT (Pre-RT), one year after RT (Post-1Y), and two years after RT (Post-2Y). Compared to Pre-RT, Post-1Y and Post-2Y demonstrated higher AS, indicating more abnormal cosmetic outcomes (Pre-RT vs. Post-1Y, P  = 0.015 and Pre-RT vs. Post-2Y, P  = 0.011). Pre-RT AS was higher in patients having major breast complications ( P  = 0.016). Patients with autologous reconstruction showed lower AS than those with tissue expander both at Pre-RT (2.00 vs. 4.19, P  = 0.008) and Post-2Y (2.89 vs. 5.00, P  = 0.010). Linear mixed effect model revealed that days after baseline were associated with increased AS ( P  = 0.007). Also, tissue expander was associated with steeper rise of AS, compared to autologous tissue ( P  = 0.015). Fractionation regimen was not associated with the change of AS ( P  = 0.389). AS detected by deep learning might be feasible in predicting cosmetic outcomes of RT-treated patients with breast reconstruction. AS should be validated in prospective studies.

Industrial Computed Tomography (CT) technology is increasingly applied in fields such as additive manufacturing and non-destructive testing, providing rich three-dimensional information for various fields, which is crucial for internal structure detection, defect detection, and product development. In subsequent processes such as analysis, simulation, and editing, three-dimensional volume data models often need to be converted into mesh models, making effective meshing of volume data essential for expanding the application scenarios and scope of industrial CT. However, the existing Marching Cubes algorithm has issues with low efficiency and poor mesh quality during the volume data meshing process. To overcome these limitations, this study proposes an innovative method for industrial CT volume data meshing based on the Iterative Smooth Signed Surface Distance (iSSD) algorithm. This method first refines the segmented voxel model, accurately extracts boundary voxels, and constructs a high-quality point cloud model. By randomly initializing the normals of the point cloud and iteratively updating the point cloud normals, the mesh is reconstructed using the SSD algorithm after each iteration update, ultimately achieving high-quality, watertight, and smooth mesh model reconstruction, ensuring the accuracy and reliability of the reconstructed mesh. Qualitative and quantitative analyses with other methods have further highlighted the excellent performance of the method proposed in this paper. This study not only improves the efficiency and quality of volume data meshing but also provides a solid foundation for subsequent three-dimensional analysis, simulation, and editing, and has important industrial application prospects and academic value.

Network traffic anomaly detection is pivotal in cybersecurity, especially as data volume grows and security requirement intensifies. This study addresses critical limitations in existing reconstruction-based methods, which quantify anomalies relying on intra-sample differences and struggle to detect drifted anomalies. In response, we propose a novel approach, the Uncertainty-Inspired Inter-Sample Differences (UnDiff) method, which leverages model uncertainty to enhance anomaly detection capabilities, particularly in scenarios involving anomaly drift. By employing evidential learning, the UnDiff model gathers evidence to minimize uncertainty in normal network traffic, enhancing its ability to differentiate between normal and anomalous traffic. To overcome the limitations of intra-sample difference quantification in reconstruction-based methods, we propose a novel anomaly score based on inter-sample uncertainty deviation that directly quantifies the anomaly degree. Benefiting from a concise model design and parameterized uncertainty quantification, UnDiff achieves high efficiency. Extensive experiments on three benchmarks demonstrate UnDiff's superior performance in detecting both undrifted and drifted anomalies with minimal computational overhead.

By pushing computing resources from the cloud to the network edge close to mobile users, mobile edge computing (MEC) enables low latency for a wide variety of applications. Nevertheless, in dynamic MEC systems, MEC services are challenged by the risks of runtime reliability anomalies. Detecting runtime reliability anomalies for MEC services is challenging yet critical to ensuring the stability of MEC systems. The effectiveness of existing anomaly detection methods suffers from poor performance when handling MEC services’ large-volume, continuous, and volatile reliability streaming data. The key is to identify significant changes in the distribution of MEC services’ current reliability streaming data compared with their historical performance. Inspired by concept drift, this paper proposes B-Detection, a boosting Long Short-Term Memory (LSTM) Autoencoder for detecting MEC services’ runtime reliability anomalies based on distribution dissimilarity evaluation. B-Detection employs a deep learning method named LSTM Autoencoder to characterize the MEC services’ historical reliability data distribution. To cope with the challenge of modeling complex distribution characteristics of MEC services’ historical reliability streaming data and guarantee the real-time performance of B-Detection, we enhance LSTM Autoencoder with a weight-based reservoir sampling technique and an LSTM boosting algorithm. The reconstruction loss of the trained LSTM Autoencoder model is estimated for the up-to-date reliability streaming data, and the result is used to infer MEC services’ runtime reliability anomalies. The performance of B-Detection is verified through a series of experiments conducted on a real-world dataset.