IEEE transactions on medical imaging
-
IEEE Trans Med Imaging · May 2016
Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning.
Remarkable progress has been made in image recognition, primarily due to the availability of large-scale annotated datasets and deep convolutional neural networks (CNNs). CNNs enable learning data-driven, highly representative, hierarchical image features from sufficient training data. However, obtaining datasets as comprehensively annotated as ImageNet in the medical imaging domain remains a challenge. ⋯ We study two specific computer-aided detection (CADe) problems, namely thoraco-abdominal lymph node (LN) detection and interstitial lung disease (ILD) classification. We achieve the state-of-the-art performance on the mediastinal LN detection, and report the first five-fold cross-validation classification results on predicting axial CT slices with ILD categories. Our extensive empirical evaluation, CNN model analysis and valuable insights can be extended to the design of high performance CAD systems for other medical imaging tasks.
-
IEEE Trans Med Imaging · Apr 2016
Theoretical Analysis of Penalized Maximum-Likelihood Patlak Parametric Image Reconstruction in Dynamic PET for Lesion Detection.
Detecting cancerous lesions is a major clinical application of emission tomography. In a previous work, we studied penalized maximum-likelihood (PML) image reconstruction for lesion detection in static PET. Here we extend our theoretical analysis of static PET reconstruction to dynamic PET. ⋯ The proposed method is validated using computer-based Monte Carlo simulations. Good agreements between the theoretical predictions and the Monte Carlo results are observed. Both theoretical predictions and Monte Carlo simulation results show the benefit of the indirect and direct methods under optimized regularization parameters in dynamic PET reconstruction for lesion detection, when compared with the conventional static PET reconstruction.
-
IEEE Trans Med Imaging · Mar 2016
Determining the Performance of Fluorescence Molecular Imaging Devices Using Traceable Working Standards With SI Units of Radiance.
To date, no emerging preclinical or clinical near-infrared fluorescence (NIRF) imaging devices for noninvasive and/or surgical guidance have their performances validated on working standards with SI units of radiance that enable comparison or quantitative quality assurance. In this work, we developed and deployed a methodology to calibrate a stable, solid phantom for emission radiance with International System of Units (SI) units of mW ·sr(-1) ·cm(-2) for use in characterizing the measurement sensitivity of ICCD and IsCMOS detection, signal-to-noise ratio, and contrast. ⋯ Contrast depended upon the camera settings (binning and integration time) and gain of intensifier. Finally, because the architecture of CMOS and CCD camera systems results in vastly different performance, we comment on the utility of these technologies for small animal imaging as well as clinical applications for noninvasive and surgical guidance.
-
IEEE Trans Med Imaging · Feb 2015
Simultaneous phase unwrapping and removal of chemical shift (SPURS) using graph cuts: application in quantitative susceptibility mapping.
Quantitative susceptibility mapping (QSM) is a magnetic resonance imaging technique that reveals tissue magnetic susceptibility. It relies on having a high quality field map, typically acquired with a relatively long echo spacing and long final TE. Applications of QSM outside the brain require the removal of fat contributions to the total signal phase. ⋯ The result from SPURS is then used as the initial guess for a voxel-wise iterative decomposition of water and fat with echo asymmetric and least-squares estimation (IDEAL). The estimated 3-D field maps are used to compute QSM in body regions outside of the brain, such as the liver. Experimental results show substantial improvements in field map estimation, water/fat separation and reconstructed QSM compared to two existing water/fat separation methods on 1.5T and 3T magnetic resonance human data with long echo spacing and rapid field map variation.
-
IEEE Trans Med Imaging · May 2014
An analysis of whole body tracer kinetics in dynamic PET studies with application to image-based blood input function extraction.
In a positron emission tomography (PET) study, the local uptake of the tracer is dependent on vascular delivery and retention. For dynamic studies the measured uptake time-course information can be best interpreted when knowledge of the time-course of tracer in the blood is available. This is certainly true for the most established tracers such as 18F-Fluorodeoxyglucose (FDG) and 15O-Water (H2O). ⋯ A collection of arterially sampled data from PET studies with FDG and H2O is used to illustrate the methodology. These data analyses are highly supportive of the overall modeling approach. An adaptation of the model to the problem of extraction of arterial blood signals from imaging data is also developed and promising preliminary results for cerebral and thoracic imaging studies with FDG and H2O are obtained.