RESUMO
BACKGROUND: Computed tomography (CT) and magnetic resonance imaging (MRI) are indicated for use in preoperative planning and may complicate diagnosis and place a burden on patients with lumbar disc herniation. PURPOSE: To investigate the diagnostic potential of MRI-based synthetic CT with conventional CT in the diagnosis of lumbar disc herniation. MATERIAL AND METHODS: After obtaining prior institutional review board approval, 19 patients who underwent conventional and synthetic CT imaging were enrolled in this prospective study. Synthetic CT images were generated from the MRI data using U-net. The two sets of images were compared and analyzed qualitatively by two musculoskeletal radiologists. The images were rated on a 4-point scale to determine their subjective quality. The agreement between the conventional and synthetic images for a diagnosis of lumbar disc herniation was determined independently using the kappa statistic. The diagnostic performances of conventional and synthetic CT images were evaluated for sensitivity, specificity, and accuracy, and the consensual results based on T2-weighted imaging were employed as the reference standard. RESULTS: The inter-reader and intra-reader agreement were almost moderate for all evaluated modalities (κ = 0.57-0.79 and 0.47-0.75, respectively). The sensitivity, specificity, and accuracy for detecting lumbar disc herniation were similar for synthetic and conventional CT images (synthetic vs. conventional, reader 1: sensitivity = 91% vs. 81%, specificity = 83% vs. 100%, accuracy = 87% vs. 91%; P < 0.001; reader 2: sensitivity = 84% vs. 81%, specificity = 85% vs. 98%, accuracy = 84% vs. 90%; P < 0.001). CONCLUSION: Synthetic CT images can be used in the diagnostics of lumbar disc herniation.
Assuntos
Deslocamento do Disco Intervertebral , Humanos , Deslocamento do Disco Intervertebral/diagnóstico por imagem , Estudos Prospectivos , Estudos de Viabilidade , Vértebras Lombares/diagnóstico por imagem , Tomografia Computadorizada por Raios X/métodos , Imageamento por Ressonância Magnética/métodosRESUMO
BACKGROUND: High breast density is a strong risk factor for breast cancer. As such, high consistency and accuracy in breast density assessment is necessary. PURPOSE: To validate our proposed deep learning (DL) model and explore its impact on radiologists on density assessments. MATERIAL AND METHODS: A total of 3732 mammographic cases were collected as a validated set: 1686 cases before the implementation of the DL model and 2046 cases after the DL model. Five radiologists were divided into two groups (junior and senior groups) to assess all mammograms using either two- or four-category evaluation. Linear-weighted kappa (K) and intraclass correlation coefficient (ICC) statistics were used to analyze the consistency between radiologists before and after implementation of the DL model. RESULTS: The accuracy and clinical acceptance of the DL model for the junior group were 96.3% and 96.8% for two-category evaluation, and 85.6% and 89.6% for four-category evaluation, respectively. For the senior group, the accuracy and clinical acceptance were 95.5% and 98.0% for two-category evaluation, and 84.3% and 95.3% for four-category evaluation, respectively. The consistency within the junior group, the senior group, and among all radiologists improved with the help of the DL model. For two-category, their K and ICC values improved to 0.81, 0.81, and 0.80 from 0.73, 0.75, and 0.76. And for four-category, their K and ICC values improved to 0.81, 0.82, and 0.82 from 0.73, 0.79, and 0.78, respectively. CONCLUSION: The DL model showed high accuracy and clinical acceptance in breast density categories. It is helpful to improve radiologists' consistency.
Assuntos
Neoplasias da Mama , Aprendizado Profundo , Feminino , Humanos , Densidade da Mama , População do Leste Asiático , Mamografia , Neoplasias da Mama/diagnóstico por imagemRESUMO
OBJECTIVES: To investigate the value of an artificial intelligence (AI) system in assisting radiologists to improve the assessment accuracy of BI-RADS 0 cases in mammograms. METHODS: We included 34,654 consecutive digital mammography studies, collected between January 2011 and January 2019, among which, 1088 cases from 1010 unique patients with initial BI-RADS 0 assessment who were recalled during 2 years of follow-up were used in this study. Two mid-level radiologists retrospectively re-assessed these BI-RADS 0 cases with the assistance of an AI system developed by us previously. In addition, four entry-level radiologists were split into two groups to cross-read 80 cases with and without the AI. Diagnostic performance was evaluated using the follow-up diagnosis or biopsy results as the reference standard. RESULTS: Of the 1088 cases, 626 were actually normal (BI-RADS 1 and no recall required). Assisted by the AI system, 351 (56%) and 362 (58%) normal cases were correctly identified by the two mid-level radiologists hence can be avoided for unnecessary follow-ups. However, they would have missed 12 (10 invasive cancers and 2 ductal carcinoma in situ cancers) and 6 (invasive cancers) malignant lesions respectively as a result. These missed lesions were not highly malignant tumors. The inter-rater reliability of entry-level radiologists increased from 0.20 to 0.30 (p < 0.005) by introducing the AI. CONCLUSION: The AI system can effectively assist mid-level radiologists in reducing unnecessary follow-ups of mammographically indeterminate breast lesions and reducing the benign biopsy rate without missing highly malignant tumors. KEY POINTS: ⢠The artificial intelligence system could assist mid-level radiologists in effectively reducing unnecessary BI-RADS 0 mammogram recalls and the benign biopsy rate without missing highly malignant tumors. ⢠The artificial intelligence system was capable of detecting low suspicion lesions from heterogeneously and extremely dense breasts that radiologists tended to miss. ⢠The use of an artificial intelligence system may improve the inter-rater reliability and sensitivity, and reduce the reading time of entry-level radiologists in assessing potential lesions in BI-RADS 0 mammograms.
Assuntos
Inteligência Artificial , Neoplasias da Mama , Neoplasias da Mama/diagnóstico por imagem , Feminino , Humanos , Mamografia , Radiologistas , Reprodutibilidade dos Testes , Estudos RetrospectivosRESUMO
As an emerging modality for whole breast imaging, ultrasound tomography (UST), has been adopted for diagnostic purposes. Efficient segmentation of an entire breast in UST images plays an important role in quantitative tissue analysis and cancer diagnosis, while major existing methods suffer from considerable time consumption and intensive user interaction. This paper explores three-dimensional GrabCut (GC3D) for breast isolation in thirty reflection (B-mode) UST volumetric images. The algorithm can be conveniently initialized by localizing points to form a polygon, which covers the potential breast region. Moreover, two other variations of GrabCut and an active contour method were compared. Algorithm performance was evaluated from volume overlap ratios ( T O , target overlap; M O , mean overlap; F P , false positive; F N , false negative) and time consumption. Experimental results indicate that GC3D considerably reduced the work load and achieved good performance ( T O = 0.84; M O = 0.91; F P = 0.006; F N = 0.16) within an average of 1.2 min per volume. Furthermore, GC3D is not only user friendly, but also robust to various inputs, suggesting its great potential to facilitate clinical applications during whole-breast UST imaging. In the near future, the implemented GC3D can be easily automated to tackle B-mode UST volumetric images acquired from the updated imaging system.
RESUMO
Nano-TiO2 photocatalysis technology has attracted wide attention because of its safety, nontoxicity and long-lasting performance. However, traditional nano-TiO2 has been greatly limited in its application because its wide band gap can only be activated by ultraviolet light (λ < 387 nm). In this paper, nano-TiO2 was prepared by self-doping method. The synthesized nano-TiO2 was a single anatase crystal type with a particle size of 10 nm and uniform size. In addition, nano-TiO2 has high stability and good dispersion. More importantly, nano-TiO2 exhibits excellent visible light (400-780 nm) activity due to the decrease of bandgap from 3.20 eV to 1.80 eV (less than 2.0 eV) and the presence of a large number of hydroxyl groups on the surface of the nanoparticles. In the antibacterial test, the antibacterial rate of both E.coli and S.aureus was close to 100 % under the irradiation of household low-power LED lamps, showing excellent antibacterial performance, indicating that the prepared nano-TiO2 has broad application prospects in the field of bactericidal and bacteriostatic.
RESUMO
BACKGROUND: Fetal spinal magnetic resonance imaging (MRI) is a prenatal routine for proper assessment of fetus development, especially when suspected spinal malformations occur while ultrasound fails to provide details. Limited by hardware, fetal spine MR images suffer from its low resolution.High-resolution MR images can directly enhance readability and improve diagnosis accuracy. Image interpolation for higher resolution is required in clinical situations, while many methods fail to preserve edge structures. Edge carries heavy structural messages of objects in visual scenes for doctors to detect suspicions, classify malformations and make correct diagnosis. Effective interpolation with well-preserved edge structures is still challenging. METHOD: In this paper, we propose an edge-directed interpolation (EDI) method and apply it on a group of fetal spine MR images to evaluate its feasibility and performance. This method takes edge messages from Canny edge detector to guide further pixel modification. First, low-resolution (LR) images of fetal spine are interpolated into high-resolution (HR) images with targeted factor by bi-linear method. Then edge information from LR and HR images is put into a twofold strategy to sharpen or soften edge structures. Finally a HR image with well-preserved edge structures is generated. The HR images obtained from proposed method are validated and compared with that from other four EDI methods. Performances are evaluated from six metrics, and subjective analysis of visual quality is based on regions of interest (ROI). RESULTS: All these five EDI methods are able to generate HR images with enriched details. From quantitative analysis of six metrics, the proposed method outperforms the other four from signal-to-noise ratio (SNR), peak signal-to-noise ratio (PSNR), structure similarity index (SSIM), feature similarity index (FSIM) and mutual information (MI) with seconds-level time consumptions (TC). Visual analysis of ROI shows that the proposed method maintains better consistency in edge structures with the original images. CONCLUSIONS: The proposed method classifies edge orientations into four categories and well preserves structures. It generates convincing HR images with fine details and is suitable in real-time situations. Iterative curvature-based interpolation (ICBI) method may result in crisper edges, while the other three methods are sensitive to noise and artifacts.
Assuntos
Feto , Processamento de Imagem Assistida por Computador/métodos , Imageamento por Ressonância Magnética/métodos , Coluna Vertebral , Feminino , Humanos , Gravidez , Controle de QualidadeRESUMO
In cross-subject fall risk classification based on plantar pressure, a challenge is that data from different subjects have significant individual information. Thus, the models with insufficient generalization ability can't perform well on new subjects, which limits their application in daily life. To solve this problem, domain adaptation methods are applied to reduce the gap between source and target domain. However, these methods focus on the distribution of the source and the target domain, but ignore the potential correlation among multiple source subjects, which deteriorates domain adaptation performance. In this paper, we proposed a novel method named domain adaptation with subject fusion (SFDA) for fall risk assessment, greatly improving the cross-subject assessment ability. Specifically, SFDA synchronously carries out source target adaptation and multiple source subject fusion by domain adversarial module to reduce source-target gap and distribution distance within source subjects of same class. Consequently, target samples can learn more task-specific features from source subjects to improve the generalization ability. Experiment results show that SFDA achieved mean accuracy of 79.17 % and 73.66 % based on two backbones in a cross-subject classification manner, outperforming the state-of-the-art methods on continuous plantar pressure dataset. This study proves the effectiveness of SFDA and provides a novel tool for implementing cross-subject and few-gait fall risk assessment.
Assuntos
Marcha , Generalização Psicológica , Humanos , Aprendizagem , Medição de Risco , Coluna VertebralRESUMO
In response to the pressing need for robust disease diagnosis from gastrointestinal tract (GIT) endoscopic images, we proposed FLATer, a fast, lightweight, and highly accurate transformer-based model. FLATer consists of a residual block, a vision transformer module, and a spatial attention block, which concurrently focuses on local features and global attention. It can leverage the capabilities of both convolutional neural networks (CNNs) and vision transformers (ViT). We decomposed the classification of endoscopic images into two subtasks: a binary classification to discern between normal and pathological images and a further multi-class classification to categorize images into specific diseases, namely ulcerative colitis, polyps, and esophagitis. FLATer has exhibited exceptional prowess in these tasks, achieving 96.4% accuracy in binary classification and 99.7% accuracy in ternary classification, surpassing most existing models. Notably, FLATer could maintain impressive performance when trained from scratch, underscoring its robustness. In addition to the high precision, FLATer boasted remarkable efficiency, reaching a notable throughput of 16.4k images per second, which positions FLATer as a compelling candidate for rapid disease identification in clinical practice.
RESUMO
Nitrous oxide (N2O) has demonstrated an antidepressant effect for treatment-resistant depression (TRD), but no studies investigated the effects of N2O on different cognition domains. This study aimed to test whether N2O would display pro-cognitive effects. We conducted a double-blinded, placebo-controlled, randomized controlled trial, 44 patients with TRD were randomized to N2O group (one-hour inhalation of 50% N2O/50% oxygen) or placebo group (50% air/50% oxygen). Thirty-four patients completed cognitive tests at the pre-treatment phase, 1-week, and 2 weeks post-treatment including subjective cognitive function, processing speed, attention, and executive function. Although the antidepressant effect of N2O was not significant at 1 week, patients still showed better performance of executive function at 1 week after receiving N2O compared with the placebo. Moreover, this significant improvement still existed at 1 week after controlling for the change in depressive symptoms over-time. Additionally, no significant difference was observed in subjective cognitive function, processing speed, and attention between these two groups across the 2-week follow-up period. As the first study investigating the treatment effects of N2O on improving cognitive function in TRD patients, the current study indicated that N2O has a potential pro-cognitive effect on executive function and this effect might be independent from improvements in depressive symptoms.
Assuntos
Transtorno Depressivo Resistente a Tratamento , Óxido Nitroso , Humanos , Óxido Nitroso/uso terapêutico , Transtorno Depressivo Resistente a Tratamento/tratamento farmacológico , Método Duplo-Cego , Antidepressivos/uso terapêutico , Oxigênio/uso terapêutico , Resultado do TratamentoRESUMO
Continuous fall risk assessment and real-time high falling risk warning are extremely necessary for the elderly, to protect their lives and ensure their quality of life. Wearable in-shoe pressure sensors have the potential to achieve these targets, due to their adequate wearing comfort. However, it is a great challenge to remove the individual differences of foot pressure data and identify the accurate fall risk from fewer gait cycles to realize real-time warning. We explored a hierarchical deep learning network named MhNet for real-time fall risk assessment, which utilized the advantages of two-layer network, to reach hierarchical tasks to reduce probability of misidentification of high fall risk subjects, by establishing a borderline category using the rehabilitation labels, and extracting multi-scale spatio-temporal features. It was trained by using a wearable plantar pressure dataset collected from 48 elderly subjects. This method could achieve a real time fall risk identification accuracy of 73.27% by using only 9 gaits, which was superior to traditional methods. Moreover, the sensitivity reached 76.72%, proving its strength in identifying high risk samples. MhNet might be a promising way in real-time fall risk assessment for the elderly in their daily activities.
Assuntos
Aprendizado Profundo , Dispositivos Eletrônicos Vestíveis , Idoso , Marcha , Humanos , Qualidade de Vida , Medição de RiscoRESUMO
In recent years, with the rapid development of machine learning, automatic emotion recognition based on electroencephalogram (EEG) signals has received increasing attention. However, owing to the great variance of EEG signals sampled from different subjects, EEG-based emotion recognition experiences the individual difference problem across subjects, which significantly hinders recognition performance. In this study, we presented a method for EEG-based emotion recognition using a combination of a multi-scale residual network (MSRN) and meta-transfer learning (MTL) strategy. The MSRN was used to represent connectivity features of EEG signals in a multi-scale manner, which utilized different receptive fields of convolution neural networks to capture the interactions of different brain regions. The MTL strategy fully used the merits of meta-learning and transfer learning to significantly reduce the gap in individual differences between various subjects. The proposed method can not only further explore the relationship between connectivity features and emotional states but also alleviate the problem of individual differences across subjects. The average cross-subject accuracies of the proposed method were 71.29% and 71.92% for the valence and arousal tasks on the DEAP dataset, respectively. It achieved an accuracy of 87.05% for the binary classification task on the SEED dataset. The results show that the framework has a positive effect on the cross-subject EEG emotion recognition task.
Assuntos
Nível de Alerta , Eletroencefalografia , Emoções , Humanos , Aprendizado de Máquina , Redes Neurais de ComputaçãoRESUMO
The high fall rate of the elderly brings enormous challenges to families and the medical system; therefore, early risk assessment and intervention are quite necessary. Compared to other sensor-based technologies, in-shoe plantar pressure sensors, effectiveness and low obtrusiveness are widely used for long-term fall risk assessments because of their portability. While frequently-used bipedal center-of-pressure (COP) features are derived from a pressure sensing platform, they are not suitable for the shoe system or pressure insole owing to the lack of relative position information. Therefore, in this study, a definition of "weak foot" was proposed to solve the sensitivity problem of single foot features and facilitate the extraction of temporal consistency related features. Forty-four multi-dimensional weak foot features based on single foot COP were correspondingly extracted; notably, the relationship between the fall risk and temporal inconsistency in the weak foot were discussed in this study, and probability distribution method was used to analyze the symmetry and temporal consistency of gait lines. Though experiments, foot pressure data were collected from 48 subjects with 24 high risk (HR) and 24 low risk (LR) ones obtained by the smart footwear system. The final models with 87.5% accuracy and 100% sensitivity on test data outperformed the base line models using bipedal COP. The results and feature space shown the novel features of wearable plantar pressure could comprehensively evaluate the difference between HR and LR groups. Our fall risk assessment models based on these features had good generalization performance, and showed practicability and reliability in real-life monitoring situations.
Assuntos
Pé , Dispositivos Eletrônicos Vestíveis , Idoso , Humanos , Pressão , Reprodutibilidade dos Testes , Medição de RiscoRESUMO
Background: Intestinal metaplasia (IM) of the gastric cardia is an important premalignant lesion. However, there is limited information concerning its epidemiological and molecular features. Herein, we aimed to provide an overview of the epidemiological data for gastric cardiac IM and evaluate the role of EYA transcriptional coactivator and phosphatase 4 (EYA4) as an epigenetic biomarker for gastric cardiac IM. Methods: The study was conducted in the context of the gastric cardiac precancerous lesion program in southern China, which included 718 non-cancer participants, who undertook endoscopic biopsy and pathological examination in three endoscopy centers, between November 2018 and November 2021. Pyrosequencing and immunohistochemistry were performed to examine the DNA methylation status and protein expression level of EYA4. Results: Gastric cardiac IM presented in 14.1% (101/718) of participants and was more common among older (>50 years; 22.0% [95% CI: 17.8-26.8]) than younger participants (≤50 years; 6.7% [95% CI: 4.5-9.9]; P < 0.001). IM was more common in male participants (16.9% [95% CI: 13.2-21.3] vs. 11.3% [95% CI: 8.3-15.1]; P = 0.04). Pyrosequencing revealed that IM tissues exhibited significantly higher DNA methylation levels in EYA4 gene than normal tissues (P = 0.016). Further, the protein expression level of EYA4 was reduced in IM and absent in intraepithelial neoplasia tissues compared to normal tissues (P < 0.001). Conclusions: Detection rates of gastric cardiac IM increase with age and are higher in men. Our findings highlight the important role of promoter hypermethylation and downregulation of EYA4 in gastric cardiac IM development.
Assuntos
Lesões Pré-Cancerosas , Gastropatias , Masculino , Humanos , Cárdia , Metilação de DNA , Metaplasia/genética , TransativadoresRESUMO
Nitrous oxide (N2O), an N-methyl-D-aspartate glutamate receptor antagonist, has demonstrated a rapid-onset antidepressant effect for treatment-resistant depression (TRD) preliminarily in the United States. This study aimed to test the efficacy and safety of N2O for TRD patients in China. In this double-blinded, placebo-controlled trial, 44 patients with TRD were randomized to receive a one-hour inhalation of a mixture of either 50% N2O/50% oxygen (N2O group) or 50% air/50% oxygen (placebo group). The primary outcome was the change on the 17-item Hamilton Depression Rating Scale (HDRS-17) over a course of two weeks. Using modified intention-to-treat analysis, the between group difference was found in the HDRS-17 score at two hours and 24 h after inhalation treatment, while no significant difference was found in week one and week two. Patients receiving N2O reported a significantly higher number of adverse events. All the adverse events lasted for no more than 24 h. No serious adverse events were reported. A single inhalation of 50% N2O effectively alleviates depression in patients with TRD in China. The efficacy lasts for no more than one week. N2O is safe for patients with TRD.
Assuntos
Transtorno Depressivo Resistente a Tratamento , Óxido Nitroso , Humanos , Óxido Nitroso/efeitos adversos , Depressão , Método Duplo-Cego , Resultado do Tratamento , Transtorno Depressivo Resistente a Tratamento/tratamento farmacológico , Oxigênio/efeitos adversosRESUMO
Recent innovations in 3D printing technologies and processes have influenced how landscape products are designed, built, and developed. In landscape architecture, reduced-size models are 3D-printed to replicate full-size structures. However, high surface roughness usually occurs on the surfaces of such 3D-printed components, which requires additional post-treatment. In this work, we develop a new type of landscape design structure based on the fused deposition modeling (FDM) technique and present a laser polishing method for FDM-fabricated polylactic acid (PLA) mechanical components, whereby the surface roughness of the laser-polished surfaces is reduced from over Ra 15 µm to less than 0.25 µm. The detailed results of thermodynamics and microstructure evolution are further analyzed during laser polishing. The stability and accuracy of the results are evaluated based on the standard deviation. Additionally, the superior tensile and flexural properties are examined in the laser-polished layer, in which the ultimate tensile strength (UTS) is increased by up to 46.6% and the flexural strength is increased by up to 74.5% compared with the as-fabricated components. Finally, a real polished landscape model is simulated and optimized using a series of scales.
RESUMO
Metal microspheres doping porous carbon (MMPC), which was prepared using in-situ pyrolysis reduction strategy, could enhance the thermal conductivity of shape-stabilized phase change material (ss-PCM) prepared by MMPC as the matrix. However, in previous studies that were reported, the preparation of MMPC needed to synthesize porous carbon by pyrolysis firstly, and then porous carbon adsorbed metal ions was pyrolyzed again to obtain MMPC, which was tedious and energy-prodigal. In this study, a one-step pyrolysis strategy was developed for the synthesis of MMPC through the pyrolyzation of wheat bran adsorbed copper ions, and the copper microspheres doping wheat bran biochar (CMS-WBB) was prepared. The CMS-WBB was taken as the supporter of stearic acid (SA) to synthesize the ss-PCM of SA/CMS-WBB. The study results about the thermal properties of SA/CMS-WBB demonstrated that the introduction of copper microspheres could not only improve the thermal conductivity of SA/CMS-WBB, but also could increase the SA loading amount of wheat bran biochar. More importantly, the CMS-WBB could be obtained by only one-step pyrolysis, which greatly simplified the preparation process and saved energy consumption. Furthermore, the raw material of wheat bran is a kind of agricultural waste, which is abundant, cheap and easy to obtain. Hence, the SA/CMS-WBB synthesized in this study had huge potentialities in thermal management applications, and a simplified method for improving the thermal properties of ss-PCMs was provided.
RESUMO
Background: The difficulty in timely evaluating patient response to antidepressants has brought great challenge to the treatment of major depressive disorder (MDD). Some studies found that the electroencephalogram (EEG) microstates might be a reliable marker to evaluate patient response to treatment. The present study aims to evaluate the relationship between EEG microstate parameters and MDD symptoms before and after treatment to identify predictive biological markers for patient response. Methods: Thirty drug-naïve MDD patients (20 females and 10 males) were enrolled in this study. All the patients received effective dosages of selective serotonin reuptake inhibitors, and EEG recordings were collected at baseline and 2 weeks of treatment. Brain activities during the eyes-closed state were recorded using 64-channel electroencephalography, and the patients' microstates were clustered into four maps according to their topography (labeled A, B, C, and D). The differences of EEG microstates before and after treatment were compared using paired t-test. Spearman correlation coefficients were calculated to identify the relationships between the improvement of depression and anxiety symptoms and microstate parameters. Results: The mean duration (69.67 ± 10.33 vs. 64.00 ± 7.70, p < 0.001) and occurrence (4.06 ± 0.69, vs. 3.69 ± 0.70, p = 0.002) of microstate B decreased significantly after treatment. The proportion of microstate B also decreased (27.53 ± 5.81, vs. 23.23 ± 4.61, p < 0.001), while the occurrence of microstate A increased after treatment. A significant negative correlation was found between the change of score of Hamilton Rating Scale for Anxiety and the increase of the occurrence of microstate A (r = -0.431, p < 0.05) after 2 weeks of treatment. The reduction of the duration of microstate B was found to be predictive of patient response to antidepressants after 3 months. Conclusion: This study explored the relationship between changes of EEG microstates and patient response to antidepressants. Depression symptoms might be associated with the duration of microstate B and anxiety symptoms related to the occurrence of microstate A. Therefore, the duration of microstate B and the occurrence of microstate A are potential biological markers for MDD patients' early response and further clinical outcomes.
RESUMO
Many existing approaches for mammogram analysis are based on single view. Some recent DNN-based multi-view approaches can perform either bilateral or ipsilateral analysis, while in practice, radiologists use both to achieve the best clinical outcome. MommiNet is the first DNN-based tri-view mass identification approach, which can simultaneously perform bilateral and ipsilateral analysis of mammographic images, and in turn, can fully emulate the radiologists' reading practice. In this paper, we present MommiNet-v2, with improved network architecture and performance. Novel high-resolution network (HRNet)-based architectures are proposed to learn the symmetry and geometry constraints, to fully aggregate the information from all views for accurate mass detection. A multi-task learning scheme is adopted to incorporate both Breast Imaging-Reporting and Data System (BI-RADS) and biopsy information to train a mass malignancy classification network. Extensive experiments have been conducted on the public DDSM (Digital Database for Screening Mammography) dataset and our in-house dataset, and state-of-the-art results have been achieved in terms of mass detection accuracy. Satisfactory mass malignancy classification result has also been obtained on our in-house dataset.
Assuntos
Neoplasias da Mama , Neoplasias da Mama/diagnóstico por imagem , Bases de Dados Factuais , Detecção Precoce de Câncer , Feminino , Humanos , MamografiaRESUMO
To align multimodal images is important for information fusion, clinical diagnosis, treatment planning, and delivery, while few methods have been dedicated to matching computerized tomography (CT) and magnetic resonance (MR) images of lumbar spine. This study proposes a coarse-to-fine registration framework to address this issue. Firstly, a pair of CT-MR images are rigidly aligned for global positioning. Then, a bending energy term is penalized into the normalized mutual information for the local deformation of soft tissues. In the end, the framework is validated on 40 pairs of CT-MR images from our in-house collection and 15 image pairs from the SpineWeb database. Experimental results show high overlapping ratio (in-house collection, vertebrae 0.97 ± 0.02, blood vessel 0.88 ± 0.07; SpineWeb, vertebrae 0.95 ± 0.03, blood vessel 0.93 ± 0.10) and low target registration error (in-house collection, ≤2.00 ± 0.62 mm; SpineWeb, ≤2.37 ± 0.76 mm) are achieved. The proposed framework concerns both the incompressibility of bone structures and the nonrigid deformation of soft tissues. It enables accurate CT-MR registration of lumbar spine images and facilitates image fusion, spine disease diagnosis, and interventional treatment delivery.
Assuntos
Algoritmos , Imageamento Tridimensional , Vértebras Lombares/diagnóstico por imagem , Imagem Multimodal , Pontos de Referência Anatômicos , Humanos , Imageamento por Ressonância Magnética , Termodinâmica , Fatores de Tempo , Tomografia Computadorizada por Raios XRESUMO
Blind image quality assessment can be modeled as feature extraction followed by score prediction. It necessitates considerable expertise and efforts to handcraft features for optimal representation of perceptual image quality. This paper addresses blind image sharpness assessment by using a shallow convolutional neural network (CNN). The network takes single feature layer to unearth intrinsic features for image sharpness representation and utilizes multilayer perceptron (MLP) to rate image quality. Different from traditional methods, CNN integrates feature extraction and score prediction into an optimization procedure and retrieves features automatically from raw images. Moreover, its prediction performance can be enhanced by replacing MLP with general regression neural network (GRNN) and support vector regression (SVR). Experiments on Gaussian blur images from LIVE-II, CSIQ, TID2008 and TID2013 demonstrate that CNN features with SVR achieves the best overall performance, indicating high correlation with human subjective judgment.