Self-attention-driven retrieval of chest CT images for COVID-19 assessment.

Numerous methods have been developed for computer-aided diagnosis (CAD) of coronavirus disease-19 (COVID-19), based on chest computed tomography (CT) images. The majority of these methods are based on deep neural networks and often act as "black boxes" that cannot easily gain the trust of medical community, whereas their result is uniformly influenced by all image regions. This work introduces a novel, self-attention-driven method for content-based image retrieval (CBIR) of chest CT images. The proposed method analyzes a query CT image and returns a classification result, as well as a list of classified images, ranked according to similarity with the query. Each CT image is accompanied by a heatmap, which is derived by gradient-weighted class activation mapping (Grad-CAM) and represents the contribution of lung tissue and lesions to COVID-19 pathology. Beyond visualization, Grad-CAM weights are employed in a self-attention mechanism, in order to strengthen the influence of the most COVID-19-related image regions on the retrieval result. Experiments on two publicly available datasets demonstrate that the binary classification accuracy obtained by means of DenseNet-201 is 81.3% and 96.4%, for COVID-CT and SARS-CoV-2 datasets, respectively, with a false negative rate which is less than 3% in both datasets. In addition, the Grad-CAM-guided CBIR framework slightly outperforms the plain CBIR in most cases, with respect to nearest neighbour (NN) and first four (FF). The proposed method could serve as a computational tool for a more transparent decision-making process that could be trusted by the medical community. In addition, the employed self-attention mechanism increases the obtained retrieval performance.

Vertebrae, IVD and spinal canal boundary extraction on MRI, utilizing CT-trained active shape models.

PURPOSE: Vertebrae, intervertebral disc (IVD) and spinal canal (SC) displacements are in the root of several spinal cord pathologies. The localization and boundary extraction of these structures, along with the quantification of their displacements, provide valuable clues for assessing each pathological condition. In this work, we propose a computational method for boundary extraction of vertebrae, IVD and SC in magnetic resonance images (MRI). METHOD: Vertebrae shape priors derived from computed tomography (CT) images are used to guide vertebrae, IVD and SC boundary extraction in MRI. This strategy is dictated by three considerations: (1) CT is the modality of choice for highlighting solid structures such as vertebrae, (2) vertebrae boundaries indirectly impose constraints on the boundaries of neighbouring structures (IVD and SC), and (3) it can be observed that edges are similarly located in CT and MR images; therefore, gradient profiles and shape priors learned by active shape models (ASMs) from CT are also valid in MRI. RESULTS: Experimental comparisons on two MR image datasets demonstrate that the proposed approach obtains segmentation results, which are comparable to the state of the art. Moreover, the adopted bimodal strategy is validated by demonstrating that CT-derived shape priors lead to more accurate boundary extraction than MRI-derived shape priors, even in the case of MR image applications. CONCLUSION: Unlike existing bimodal methods, the proposed one is not dependent on the availability of CT/MR image pairs, which are not usually acquired from the same patient. In addition, unlike state-of-the-art deep learning-based methods, it is not dependent on large amounts of training data. The proposed method requires a limited amount of user intervention.

Quantitative 3-D analysis of GFAP labeled astrocytes from fluorescence confocal images.

BACKGROUND: There is a need for effective computational methods for quantifying the three-dimensional (3-D) spatial distribution, cellular arbor morphologies, and the morphological diversity of brain astrocytes to support quantitative studies of astrocytes in health, injury, and disease. NEW METHOD: Confocal fluorescence microscopy of multiplex-labeled (GFAP, DAPI) brain tissue is used to perform imaging of astrocytes in their tissue context. The proposed computational method identifies the astrocyte cell nuclei, and reconstructs their arbors using a local priority based parallel (LPP) tracing algorithm. Quantitative arbor measurements are extracted using Scorcioni's L-measure, and profiled by unsupervised harmonic co-clustering to reveal the morphological diversity. RESULTS: The proposed method identifies astrocyte nuclei, generates 3-D reconstructions of their arbors, and extracts quantitative arbor measurements, enabling a morphological grouping of the cell population. COMPARISON WITH EXISTING METHODS: Our method enables comprehensive spatial and morphological profiling of astrocyte populations in brain tissue for the first time, and overcomes limitations of prior methods. Visual proofreading of the results indicate a >95% accuracy in identifying astrocyte nuclei. The arbor reconstructions exhibited 3.2% fewer erroneous jumps in tracing, and 17.7% fewer false segments compared to the widely used fast-marching method that resulted in 9% jumps and 20.8% false segments. CONCLUSIONS: The proposed method can be used for large-scale quantitative studies of brain astrocyte distribution and morphology.

Self-parameterized active contours based on regional edge structure for medical image segmentation.

This work introduces a novel framework for unsupervised parameterization of region-based active contour regularization and data fidelity terms, which is applied for medical image segmentation. The work aims to relieve MDs from the laborious, time-consuming task of empirical parameterization and bolster the objectivity of the segmentation results. The proposed framework is inspired by an observed isomorphism between the eigenvalues of structure tensors and active contour parameters. Both may act as descriptors of the orientation coherence in regions containing edges. The experimental results demonstrate that the proposed framework maintains a high segmentation quality without the need of trial-and-error parameter adjustment.

Automated adjustment of region-based active contour parameters using local image geometry.

A principled method for active contour (AC) parameterization remains a challenging issue in segmentation research, with a potential impact on the quality, objectivity, and robustness of the segmentation results. This paper introduces a novel framework for automated adjustment of region-based AC regularization and data fidelity parameters. Motivated by an isomorphism between the weighting factors of AC energy terms and the eigenvalues of structure tensors, we encode local geometry information by mining the orientation coherence in edge regions. In this light, the AC is repelled from regions of randomly oriented edges and guided toward structured edge regions. Experiments are performed on four state-of-the-art AC models, which are automatically adjusted and applied on benchmark datasets of natural, textured and biomedical images and two image restoration models. The experimental results demonstrate that the obtained segmentation quality is comparable to the one obtained by empirical parameter adjustment, without the cumbersome and time-consuming process of trial and error.

A computer-based technique for automated spot detection in proteomics images.

This paper introduces a novel computer-based technique for automated detection of protein spots in proteomics images. The proposed technique is based on the localization of regional intensity maxima associated with protein spots and is formulated so as to ignore rectangular-shaped streaks, minimize the detection of false negatives, and allow the detection of multiple overlapping spots. Regional intensity constraints are imposed on the localized maxima in order to cope with the presence of noise and artifacts. The experimental evaluation of the proposed technique on real proteomics images demonstrates that it: 1) achieves a predictive value ( PV) and detection sensitivity (DS ) which exceed 90%; 2) outperforms Melanie software package in terms of PV , specificity, and DS; 3) ignores artifacts; 4) distinguishes multiple overlapping spots; 5) locates spots within streaks; and 6) is automated and efficient.

A computer-aided system for malignancy risk assessment of nodules in thyroid US images based on boundary features.

In this paper, a novel computer-based approach is proposed for malignancy risk assessment of thyroid nodules in ultrasound images. The proposed approach is based on boundary features and is motivated by the correlation which has been addressed in medical literature between nodule boundary irregularity and malignancy risk. In addition, local echogenicity variance is utilized so as to incorporate information associated with local echogenicity distribution within nodule boundary neighborhood. Such information is valuable for the discrimination of high-risk nodules with blurred boundaries from medium-risk nodules with regular boundaries. Analysis of variance is performed, indicating that each boundary feature under study provides statistically significant information for the discrimination of thyroid nodules in ultrasound images, in terms of malignancy risk. k-nearest neighbor and support vector machine classifiers are employed for the classification tasks, utilizing feature vectors derived from all combinations of features under study. The classification results are evaluated with the use of the receiver operating characteristic. It is derived that the proposed approach is capable of discriminating between medium-risk and high-risk nodules, obtaining an area under curve, which reaches 0.95.

Active contours guided by echogenicity and texture for delineation of thyroid nodules in ultrasound images.

Thyroid nodules are solid or cystic lumps formed in the thyroid gland and may be caused by a variety of thyroid disorders. This paper presents a novel active contour model for precise delineation of thyroid nodules of various shapes according to their echogenicity and texture, as displayed in ultrasound (US) images. The proposed model, named joint echogenicity-texture (JET), is based on a modified Mumford-Shah functional that, in addition to regional image intensity, incorporates statistical texture information encoded by feature distributions. The distributions are aggregated within the functional through new log-likelihood goodness-of-fit terms. The JET model requires only a rough region of interest within the thyroid gland as input and automatically proceeds with precise delineation of the nodules, revealing their shape and size. The performance of the JET model was validated on a range of US images displaying hypoechoic and isoechoic nodules of various shapes. The quantification of the results shows that the JET model: 1) provides precise delineations of thyroid nodules as compared to "ground truth" delineations obtained by experts and 2) copes with the limitations of the previous thyroid US delineation approaches as it is capable of delineating thyroid nodules regardless of their echogenicity or shape.

Variable background active contour model for computer-aided delineation of nodules in thyroid ultrasound images.

This paper presents a computer-aided approach for nodule delineation in thyroid ultrasound (US) images. The developed algorithm is based on a novel active contour model, named variable background active contour (VBAC), and incorporates the advantages of the level set region-based active contour without edges (ACWE) model, offering noise robustness and the ability to delineate multiple nodules. Unlike the classic active contour models that are sensitive in the presence of intensity inhomogeneities, the proposed VBAC model considers information of variable background regions. VBAC has been evaluated on synthetic images, as well as on real thyroid US images. From the quantification of the results, two major impacts have been derived: 1) higher average accuracy in the delineation of hypoechoic thyroid nodules, which exceeds 91%; and 2) faster convergence when compared with the ACWE model.