Few-Shot Wideband Tympanometry Classification in Otosclerosis via Domain Adaptation with Gaussian Processes

Otosclerosis is a common middle ear disease that requires a combination of examinations for its diagnosis in routine. In a previous study, we showed that this disease could be potentially diagnosed by wideband tympanometry (WBT) coupled with a convolutional neural network (CNN) in a rapid and non-invasive manner. We showed that deep transfer learning with data augmentation could be applied successfully on such a task. However, the involved synthetic and realistic data have a significant discrepancy that impedes the performance of transfer learning. To address this issue, a Gaussian processes-guided domain adaptation (GPGDA) algorithm was developed. It leveraged both the loss about the distribution distance calculated by the Gaussian processes and the loss of conventional cross entropy during the transferring. On a WBT dataset including 80 otosclerosis and 55 control samples, it achieved an area-under-the-curve of 97.9±1.1 percent after receiver operating characteristic analysis and an F1-score of 95.7±0.9 percent that were superior to the baseline methods (r=10, p<0.05, ANOVA). To understand the algorithm’s behavior, the role of each component in the GPGDA was experimentally explored on the dataset. In conclusion, our GPGDA algorithm appears to be an effective tool to enhance CNN-based WBT classification in otosclerosis using just a limited number of realistic data samples.

Download Full-text

Real-time monitoring hypoxia at high altitudes using electrical bioimpedance technique: an animal experiment

Journal of Applied Physiology ◽

10.1152/japplphysiol.00712.2020 ◽

2020 ◽

Author(s):

Lin Yang ◽

Meng Dai ◽

Qinglin Cao ◽

Shuai Ding ◽

Zhanqi Zhao ◽

...

Keyword(s):

Real Time ◽

Measurement Data ◽

Area Under The Curve ◽

Hypoxic Condition ◽

Control Group ◽

Characteristic Analysis ◽

Non Invasive ◽

Electrical Bioimpedance ◽

High Concordance ◽

Sensitivity Specificity

Hypoxia poses a serious threat to pilots. The aim of the study was to examine the efficacy of electrical bioimpedance (EBI) in detecting the onset of hypoxia in real time in a rabbit hypoxia model. Thirty-two New Zealand rabbits were divided equally into four groups (control group and 3 hypoxia groups, i.e. mild, moderate and severe). Hypoxia was induced by simulating various altitudes in the hypobaric oxygen chamber (3000 m, 5000 m and 8000 m). Both cerebral impedance and blood oxygen (SaO2) were monitored continuously. Results showed that the cerebral impedance increased immediately during the period of increasing altitude and decreased quickly to the initial baseline at the phase of descending altitude. Moreover, the change of cerebral impedance in mild hypoxia group (3000 m) is significantly smaller than those in the other two groups (5000 m and 8000 m, P<0.05). The changes of cerebral impedance and SaO2 were significantly correlated based on the total of measurement data (R2=0.628, P<0.001). Further, the agreement analysis performed with Bland-Altman and standardized residual plots exhibited high concordance between cerebral impedance and SaO2. Receiver operator characteristic analysis manifested that the sensitivity, specificity and area under the curve using cerebral impedance for changes of SaO2 >10% were 0.735, 0.826 and 0.845, respectively. These findings demonstrated that EBI could sensitively and accurately monitor changes of cerebral impedance induced by hypoxia, which might provide a potential tool for the real-time and non-invasive monitoring of hypoxic condition of pilots in flight for early identification of hypoxia.

Download Full-text

Ten Fast Transfer Learning Models for Carotid Ultrasound Plaque Tissue Characterization in Augmentation Framework Embedded with Heatmaps for Stroke Risk Stratification

Diagnostics ◽

10.3390/diagnostics11112109 ◽

2021 ◽

Vol 11 (11) ◽

pp. 2109

Author(s):

Skandha S. Sanagala ◽

Andrew Nicolaides ◽

Suneet K. Gupta ◽

Vijaya K. Koppula ◽

Luca Saba ◽

...

Keyword(s):

Deep Learning ◽

Transfer Learning ◽

Tissue Characterization ◽

Area Under The Curve ◽

Learning Models ◽

Non Invasive ◽

Lower Accuracy ◽

Improved Performance ◽

Higher Order Spectra

Background and Purpose: Only 1–2% of the internal carotid artery asymptomatic plaques are unstable as a result of >80% stenosis. Thus, unnecessary efforts can be saved if these plaques can be characterized and classified into symptomatic and asymptomatic using non-invasive B-mode ultrasound. Earlier plaque tissue characterization (PTC) methods were machine learning (ML)-based, which used hand-crafted features that yielded lower accuracy and unreliability. The proposed study shows the role of transfer learning (TL)-based deep learning models for PTC. Methods: As pertained weights were used in the supercomputer framework, we hypothesize that transfer learning (TL) provides improved performance compared with deep learning. We applied 11 kinds of artificial intelligence (AI) models, 10 of them were augmented and optimized using TL approaches—a class of Atheromatic™ 2.0 TL (AtheroPoint™, Roseville, CA, USA) that consisted of (i–ii) Visual Geometric Group-16, 19 (VGG16, 19); (iii) Inception V3 (IV3); (iv–v) DenseNet121, 169; (vi) XceptionNet; (vii) ResNet50; (viii) MobileNet; (ix) AlexNet; (x) SqueezeNet; and one DL-based (xi) SuriNet-derived from UNet. We benchmark 11 AI models against our earlier deep convolutional neural network (DCNN) model. Results: The best performing TL was MobileNet, with accuracy and area-under-the-curve (AUC) pairs of 96.10 ± 3% and 0.961 (p < 0.0001), respectively. In DL, DCNN was comparable to SuriNet, with an accuracy of 95.66% and 92.7 ± 5.66%, and an AUC of 0.956 (p < 0.0001) and 0.927 (p < 0.0001), respectively. We validated the performance of the AI architectures with established biomarkers such as greyscale median (GSM), fractal dimension (FD), higher-order spectra (HOS), and visual heatmaps. We benchmarked against previously developed Atheromatic™ 1.0 ML and showed an improvement of 12.9%. Conclusions: TL is a powerful AI tool for PTC into symptomatic and asymptomatic plaques.

Download Full-text

Adaptive Inception Based on Transfer Learning for Effective Visual Recognition

International Journal of Intelligent Engineering and Systems ◽

10.22266/ijies2020.1231.01 ◽

2020 ◽

Vol 13 (6) ◽

pp. 1-10

Author(s):

Balaji Sreenivasulu ◽

◽

Anjaneyulu Pasala ◽

Gaikwad Vasanth ◽

◽

...

Keyword(s):

Transfer Learning ◽

Visual Recognition ◽

Data Augmentation ◽

Research Study ◽

Domain Adaptation ◽

Machine Learning Techniques ◽

Support Vector ◽

Convolutional Networks ◽

Learning Techniques ◽

Two Phases

In computer vision, domain adaptation or transfer learning plays an important role because it learns a target classifier characteristics using labeled data from various distribution. The existing researches mostly focused on minimizing the time complexity of neural networks and it effectively worked on low-level features. However, the existing method failed to concentrate on data augmentation time and cost of labeled data. Moreover, machine learning techniques face difficulty to obtain the large amount of distributed label data. In this research study, the pre-trained network called inception layer is fine-tuned with the augmented data. There are two phases present in this study, where the effectiveness of data augmentation for Inception pre-trained networks is investigated in the first phase. The transfer learning approach is used to enhance the results of the first phase and the Support Vector Machine (SVM) is used to learn all the features extracted from inception layers. The experiments are conducted on a publicly available dataset to estimate the effectiveness of proposed method. The results stated that the proposed method achieved 95.23% accuracy, where the existing techniques namely deep neural network and traditional convolutional networks achieved 87.32% and 91.32% accuracy respectively. This validation results proved that the developed method nearly achieved 4-8% improvement in accuracy than existing techniques.

Download Full-text

Computer Tomography in the Diagnosis of Ascites: The Role of Fluid Attenuation Values

Current Medical Imaging Formerly Current Medical Imaging Reviews ◽

10.2174/1573405616666200722152105 ◽

2020 ◽

Vol 16 ◽

Author(s):

Paul-Andrei Ștefan ◽

Andrei Lebovici ◽

Csaba Csutak ◽

Carmen Mihaela Mihu

Keyword(s):

Computer Tomography ◽

Tumor Staging ◽

Area Under The Curve ◽

Characteristic Analysis ◽

Non Invasive ◽

Quantitative Measurements ◽

Fluid Collections ◽

Attenuation Values

Background: Intraperitoneal fluid accumulations are a common matter in current clinical practice, being encountered by most medical and surgical fields. Objective: To assess ascites fluid with attenuation values in the form of Hounsfield units (HU) in order to determine a non-invasive differentiation criterion for the diagnosis of intraperitoneal collections. Method: Sixty patients with known intra-peritoneal collections who underwent computer tomography (CT) for reasons such as tumor staging, post-surgical follow-up or other indications, were retrospectively included in this study. All subjects had a final pathological analysis of the fluid collections. Two radiologists measured the attenuation values for each collection. The averaged values were used for comparing benign and malignancy-related ascites (MRA), bland and hemorrhagic ascites and infected and noninfected fluid collections by consuming the Mann–Whitney U test. Also, the receiver operating characteristic analysis was performed for the statistically significant results (P<0.05), and the area under the curve (AUC) was calculated. Results: Attenuation values could differenti ate between benign and MRA (P=0.04; AUC=0.656; sensitivity, 65.52%; specificity, 71.43%) but failed to distinguish between bland ascites and ascites with hemorrhagic component (P=0.85), and between infected and noninfected fluid collections (P=0.47). Conclusion: Although the results are statistically significant, the substrate of differentiation between benign and MRA ascites cannot be clearly stated. As being the first study to investigate this issue, it opens the way for other researches in the field to determine the dynamics of imaging quantitative measurements according to the fluid’s pathological features.

Download Full-text

A generative flow-based model for volumetric data augmentation in 3D deep learning for computed tomographic colonography

International Journal of Computer Assisted Radiology and Surgery ◽

10.1007/s11548-020-02275-z ◽

2020 ◽

Author(s):

Tomoki Uemura ◽

Janne J. Näppi ◽

Yasuji Ryu ◽

Chinatsu Watari ◽

Tohru Kamiya ◽

...

Keyword(s):

Deep Learning ◽

Ct Colonography ◽

Data Augmentation ◽

Area Under The Curve ◽

Colorectal Polyps ◽

Generative Model ◽

Training Dataset ◽

Human Observer ◽

Characteristic Analysis ◽

Computed Tomographic

Abstract Purpose Deep learning can be used for improving the performance of computer-aided detection (CADe) in various medical imaging tasks. However, in computed tomographic (CT) colonography, the performance is limited by the relatively small size and the variety of the available training datasets. Our purpose in this study was to develop and evaluate a flow-based generative model for performing 3D data augmentation of colorectal polyps for effective training of deep learning in CADe for CT colonography. Methods We developed a 3D-convolutional neural network (3D CNN) based on a flow-based generative model (3D Glow) for generating synthetic volumes of interest (VOIs) that has characteristics similar to those of the VOIs of its training dataset. The 3D Glow was trained to generate synthetic VOIs of polyps by use of our clinical CT colonography case collection. The evaluation was performed by use of a human observer study with three observers and by use of a CADe-based polyp classification study with a 3D DenseNet. Results The area-under-the-curve values of the receiver operating characteristic analysis of the three observers were not statistically significantly different in distinguishing between real polyps and synthetic polyps. When trained with data augmentation by 3D Glow, the 3D DenseNet yielded a statistically significantly higher polyp classification performance than when it was trained with alternative augmentation methods. Conclusion The 3D Glow-generated synthetic polyps are visually indistinguishable from real colorectal polyps. Their application to data augmentation can substantially improve the performance of 3D CNNs in CADe for CT colonography. Thus, 3D Glow is a promising method for improving the performance of deep learning in CADe for CT colonography.

Download Full-text

Neonatal Jaundice Diagnosis Using a Smartphone Camera Based on Eye, Skin, and Fused Features with Transfer Learning

Sensors ◽

10.3390/s21217038 ◽

2021 ◽

Vol 21 (21) ◽

pp. 7038

Author(s):

Alhanoof Althnian ◽

Nada Almanea ◽

Nourah Aloboud

Keyword(s):

Machine Learning ◽

Transfer Learning ◽

Neonatal Jaundice ◽

Area Under The Curve ◽

Learning Model ◽

Diagnostic Tools ◽

Support Vector ◽

Non Invasive ◽

Periodic Monitoring ◽

Mlp Model

Neonatal jaundice is a common condition worldwide. Failure of timely diagnosis and treatment can lead to death or brain injury. Current diagnostic approaches include a painful and time-consuming invasive blood test and non-invasive tests using costly transcutaneous bilirubinometers. Since periodic monitoring is crucial, multiple efforts have been made to develop non-invasive diagnostic tools using a smartphone camera. However, existing works rely either on skin or eye images using statistical or traditional machine learning methods. In this paper, we adopt a deep transfer learning approach based on eye, skin, and fused images. We also trained well-known traditional machine learning models, including multi-layer perceptron (MLP), support vector machine (SVM), decision tree (DT), and random forest (RF), and compared their performance with that of the transfer learning model. We collected our dataset using a smartphone camera. Moreover, unlike most of the existing contributions, we report accuracy, precision, recall, f-score, and area under the curve (AUC) for all the experiments and analyzed their significance statistically. Our results indicate that the transfer learning model performed the best with skin images, while traditional models achieved the best performance with eyes and fused features. Further, we found that the transfer learning model with skin features performed comparably to the MLP model with eye features.

Download Full-text

Data augmentation for computed tomography angiography via synthetic image generation and neural domain adaptation

Current Directions in Biomedical Engineering ◽

10.1515/cdbme-2020-0015 ◽

2020 ◽

Vol 6 (1) ◽

Author(s):

Malte Seemann ◽

Lennart Bargsten ◽

Alexander Schlaefer

Keyword(s):

Computed Tomography ◽

Neural Networks ◽

Deep Learning ◽

Medical Imaging ◽

Computed Tomography Angiography ◽

Data Augmentation ◽

Domain Adaptation ◽

Synthetic Image ◽

Wide Range ◽

The Impact

AbstractDeep learning methods produce promising results when applied to a wide range of medical imaging tasks, including segmentation of artery lumen in computed tomography angiography (CTA) data. However, to perform sufficiently, neural networks have to be trained on large amounts of high quality annotated data. In the realm of medical imaging, annotations are not only quite scarce but also often not entirely reliable. To tackle both challenges, we developed a two-step approach for generating realistic synthetic CTA data for the purpose of data augmentation. In the first step moderately realistic images are generated in a purely numerical fashion. In the second step these images are improved by applying neural domain adaptation. We evaluated the impact of synthetic data on lumen segmentation via convolutional neural networks (CNNs) by comparing resulting performances. Improvements of up to 5% in terms of Dice coefficient and 20% for Hausdorff distance represent a proof of concept that the proposed augmentation procedure can be used to enhance deep learning-based segmentation for artery lumen in CTA images.

Download Full-text

Measles Virus IgG Avidity Assay for Use in Classification of Measles Vaccine Failure in Measles Elimination Settings

Clinical and Vaccine Immunology ◽

10.1128/cvi.00406-12 ◽

2012 ◽

Vol 19 (11) ◽

pp. 1810-1817 ◽

Cited By ~ 38

Author(s):

Sara Mercader ◽

Philip Garcia ◽

William J. Bellini

Keyword(s):

Measles Virus ◽

Area Under The Curve ◽

Epidemiological Data ◽

Diagnostic Tools ◽

High Avidity ◽

Characteristic Analysis ◽

Vaccine Failure ◽

Avidity Assay ◽

Igg Avidity

ABSTRACTIn regions where endemic measles virus has been eliminated, diagnostic assays are needed to assist in correctly classifying measles cases irrespective of vaccination status. A measles IgG avidity assay was configured using a commercially available measles-specific IgG enzyme immunoassay by modifying the protocol to include three 5-min washes with diethylamine (60 mM; pH 10.25) following serum incubation; serum was serially diluted, and the results were expressed as the end titer avidity index. Receiver operating characteristic analysis was used for evaluation and validation and to establish low (≤30%) and high (≥70%) end titer avidity thresholds. Analysis of 319 serum specimens expected to contain either high- or low-avidity antibodies according to clinical and epidemiological data indicated that the assay is highly accurate, with an area under the curve of 0.998 (95% confidence interval [CI], 0.978 to 1.000), sensitivity of 91.9% (95% CI, 83.2% to 97.0%), and specificity of 98.4% (95% CI, 91.6% to 100%). The assay is rapid (<2 h) and precise (standard deviation [SD], 4% to 7%). In 18 samples from an elimination setting outbreak, the assay identified 2 acute measles cases with low-avidity results; both were IgM-positive samples. Additionally, 11 patients (15 samples) with modified measles who were found to have high-avidity IgG results were classified as secondary vaccine failures; one sample with an intermediate-avidity result was not interpretable. In elimination settings, measles IgG avidity assays can complement existing diagnostic tools in confirming unvaccinated acute cases and, in conjunction with adequate clinical and epidemiologic investigation, aid in the classification of vaccine failure cases.

Download Full-text

Classification of Space Objects by Using Deep Learning with Micro-Doppler Signature Images

Sensors ◽

10.3390/s21134365 ◽

2021 ◽

Vol 21 (13) ◽

pp. 4365

Author(s):

Kwangyong Jung ◽

Jae-In Lee ◽

Nammoon Kim ◽

Sunjin Oh ◽

Dong-Wook Seo

Keyword(s):

Transfer Learning ◽

Data Augmentation ◽

Doppler Frequency ◽

Extraction Methods ◽

Classification Performance ◽

Reconstruction Method ◽

Generative Adversarial Network ◽

Adversarial Network ◽

Space Objects ◽

Target Characteristics

Radar target classification is an important task in the missile defense system. State-of-the-art studies using micro-doppler frequency have been conducted to classify the space object targets. However, existing studies rely highly on feature extraction methods. Therefore, the generalization performance of the classifier is limited and there is room for improvement. Recently, to improve the classification performance, the popular approaches are to build a convolutional neural network (CNN) architecture with the help of transfer learning and use the generative adversarial network (GAN) to increase the training datasets. However, these methods still have drawbacks. First, they use only one feature to train the network. Therefore, the existing methods cannot guarantee that the classifier learns more robust target characteristics. Second, it is difficult to obtain large amounts of data that accurately mimic real-world target features by performing data augmentation via GAN instead of simulation. To mitigate the above problem, we propose a transfer learning-based parallel network with the spectrogram and the cadence velocity diagram (CVD) as the inputs. In addition, we obtain an EM simulation-based dataset. The radar-received signal is simulated according to a variety of dynamics using the concept of shooting and bouncing rays with relative aspect angles rather than the scattering center reconstruction method. Our proposed model is evaluated on our generated dataset. The proposed method achieved about 0.01 to 0.39% higher accuracy than the pre-trained networks with a single input feature.

Download Full-text

A [68Ga]Ga-DOTANOC PET/CT Radiomic Model for Non-Invasive Prediction of Tumour Grade in Pancreatic Neuroendocrine Tumours

Diagnostics ◽

10.3390/diagnostics11050870 ◽

2021 ◽

Vol 11 (5) ◽

pp. 870

Author(s):

Alessandro Bevilacqua ◽

Diletta Calabrò ◽

Silvia Malavasi ◽

Claudio Ricci ◽

Riccardo Casadei ◽

...

Keyword(s):

Neuroendocrine Tumour ◽

Area Under The Curve ◽

Tumour Volume ◽

Assessment Method ◽

Second Order ◽

Tumour Grade ◽

Non Invasive ◽

Pet Ct ◽

Median Test ◽

Fdg Pet Ct

Predicting grade 1 (G1) and 2 (G2) primary pancreatic neuroendocrine tumour (panNET) is crucial to foresee panNET clinical behaviour. Fifty-one patients with G1-G2 primary panNET demonstrated by pre-surgical [68Ga]Ga-DOTANOC PET/CT and diagnostic conventional imaging were grouped according to the tumour grade assessment method: histology on the whole excised primary lesion (HS) or biopsy (BS). First-order and second-order radiomic features (RFs) were computed from SUV maps for the whole tumour volume on HS. The RFs showing the lowest p-values and the highest area under the curve (AUC) were selected. Three radiomic models were assessed: A (trained on HS, validated on BS), B (trained on BS, validated on HS), and C (using the cross-validation on the whole dataset). The second-order normalized homogeneity and entropy was the most effective RFs couple predicting G2 and G1. The best performance was achieved by model A (test AUC = 0.90, sensitivity = 0.88, specificity = 0.89), followed by model C (median test AUC = 0.87, sensitivity = 0.83, specificity = 0.82). Model B performed worse. Using HS to train a radiomic model leads to the best prediction, although a “hybrid” (HS+BS) population performs better than biopsy-only. The non-invasive prediction of panNET grading may be especially useful in lesions not amenable to biopsy while [68Ga]Ga-DOTANOC heterogeneity might recommend FDG PET/CT.

Download Full-text