IEEE transactions on medical imaging
-
IEEE Trans Med Imaging · Aug 2020
Dual-Sampling Attention Network for Diagnosis of COVID-19 From Community Acquired Pneumonia.
The coronavirus disease (COVID-19) is rapidly spreading all over the world, and has infected more than 1,436,000 people in more than 200 countries and territories as of April 9, 2020. Detecting COVID-19 at early stage is essential to deliver proper healthcare to the patients and also to protect the uninfected population. To this end, we develop a dual-sampling attention network to automatically diagnose COVID-19 from the community acquired pneumonia (CAP) in chest computed tomography (CT). ⋯ In the testing stage, we employ another independent large-scale testing dataset including 2796 CT scans from 2057 patients. Results show that our algorithm can identify the COVID-19 images with the area under the receiver operating characteristic curve (AUC) value of 0.944, accuracy of 87.5%, sensitivity of 86.9%, specificity of 90.1%, and F1-score of 82.0%. With this performance, the proposed algorithm could potentially aid radiologists with COVID-19 diagnosis from CAP, especially in the early stage of the COVID-19 outbreak.
-
IEEE Trans Med Imaging · Jul 2020
Unsupervised Domain Adaptation to Classify Medical Images Using Zero-Bias Convolutional Auto-Encoders and Context-Based Feature Augmentation.
The accuracy and robustness of image classification with supervised deep learning are dependent on the availability of large-scale labelled training data. In medical imaging, these large labelled datasets are sparse, mainly related to the complexity in manual annotation. Deep convolutional neural networks (CNNs), with transferable knowledge, have been employed as a solution to limited annotated data through: 1) fine-tuning generic knowledge with a relatively smaller amount of labelled medical imaging data, and 2) learning image representation that is invariant to different domains. ⋯ We also propose a context-based feature augmentation scheme to improve the discriminative power of the feature representation. We evaluated our approach on 3 public medical image datasets and compared it to other state-of-the-art supervised CNNs. Our unsupervised approach achieved better accuracy when compared to other conventional unsupervised methods and baseline fine-tuned CNNs.
-
IEEE Trans Med Imaging · Jun 2020
4D Functional Imaging of the Rat Brain Using a Large Aperture Row-Column Array.
Functional ultrasound imaging (fUS) recently emerged as a promising neuroimaging modality to image and monitor brain activity based on cerebral blood volume response (CBV) and neurovascular coupling. fUS offers very good spatial and temporal resolutions compared to fMRI gold standard as well as simplicity and portability. It was recently extended to 4D fUS imaging in preclinical settings although this approach remains limited and complex. Indeed 4D fUS requires a 2D matrix probe and specific hardware able to drive the N2 elements of the probe with thousands of electronic channels. ⋯ Doppler volumes of the whole rat brain were obtained in vivo at high rates (23 dB CNR at 156 Hz and 19 dB CNR at 313 Hz). Visual and whiskers stimulations were performed and the corresponding CBV increases were reconstructed in 3D with successful functional activation detected in the superior colliculus and somato-sensorial cortex respectively. This proof of concept study demonstrates for the first time the use of a low-channel count RCA array for in vivo 4D fUS imaging in the whole rat brain.
-
IEEE Trans Med Imaging · Apr 2020
Comparative StudyComparison of NIR Versus SWIR Fluorescence Image Device Performance Using Working Standards Calibrated With SI Units.
Recently, fluorescence imaging using shortwave infrared light (SWIR, 1,000-2,000 nm) has been proposed as having advantage over conventional near-infrared fluorescence (NIRF) imaging due to the reduced tissue scattering, negligible autofluorescence, comparable tissue absorption, and the discovery that indocyanine green (ICG), used clinically as a NIRF contrast agent, also has fluorescence emission in SWIR regime. Images of ICG in small animals acquired by commercial Si-based and InGaAs-based imaging cameras have been qualitatively compared, however the lack of working standards to quantify performance of these imaging systems limits quantitative comparison. ⋯ In addition, we present small animal and large animal imaging with ICG for qualitative comparison of the same SWIR fluorescence and NIRF imaging systems. Results suggest that SWIR fluorescence imaging of ICG may have superior resolution in small animal imaging compared to NIRF imaging, but lack of measurement sensitivity, SNR, contrast, as well as water absorption limits deep penetration in large animals.
-
IEEE Trans Med Imaging · Aug 2019
Lung and Pancreatic Tumor Characterization in the Deep Learning Era: Novel Supervised and Unsupervised Learning Approaches.
Risk stratification (characterization) of tumors from radiology images can be more accurate and faster with computer-aided diagnosis (CAD) tools. Tumor characterization through such tools can also enable non-invasive cancer staging, prognosis, and foster personalized treatment planning as a part of precision medicine. In this papet, we propose both supervised and unsupervised machine learning strategies to improve tumor characterization. ⋯ Inspired by learning from label proportion approaches in computer vision, we propose to use proportion-support vector machine for characterizing tumors. We also seek the answer to the fundamental question about the goodness of "deep features" for unsupervised tumor classification. We evaluate our proposed supervised and unsupervised learning algorithms on two different tumor diagnosis challenges: lung and pancreas with 1018 CT and 171 MRI scans, respectively, and obtain the state-of-the-art sensitivity and specificity results in both problems.