Chung-Ming Chen

National Taiwan University, T’ai-pei, Taipei, Taiwan

Are you Chung-Ming Chen?

Claim your profile

Publications (57)100.99 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Background This paper proposes a semantic segmentation algorithm that provides the spatial distribution patterns of pulmonary ground-glass nodules with solid portions in computed tomography (CT) images. Methods The proposed segmentation algorithm, anatomy packing with hierarchical segments (APHS), performs pulmonary nodule segmentation and quantification in CT images. In particular, the APHS algorithm consists of two essential processes: hierarchical segmentation tree construction and anatomy packing. It constructs the hierarchical segmentation tree based on region attributes and local contour cues along the region boundaries. Each node of the tree corresponds to the soft boundary associated with a family of nested segmentations through different scales applied by a hierarchical segmentation operator that is used to decompose the image in a structurally coherent manner. The anatomy packing process detects and localizes individual object instances by optimizing a hierarchical conditional random field model. Ninety-two histopathologically confirmed pulmonary nodules were used to evaluate the performance of the proposed APHS algorithm. Further, a comparative study was conducted with two conventional multi-label image segmentation algorithms based on four assessment metrics: the modified Williams index, percentage statistic, overlapping ratio, and difference ratio. Results Under the same framework, the proposed APHS algorithm was applied to two clinical applications: multi-label segmentation of nodules with a solid portion and surrounding tissues and pulmonary nodule segmentation. The results obtained indicate that the APHS-generated boundaries are comparable to manual delineations with a modified Williams index of 1.013. Further, the resulting segmentation of the APHS algorithm is also better than that achieved by two conventional multi-label image segmentation algorithms. Conclusions The proposed two-level hierarchical segmentation algorithm effectively labelled the pulmonary nodule and its surrounding anatomic structures in lung CT images. This suggests that the generated multi-label structures can potentially serve as the basis for developing related clinical applications.
    BioMedical Engineering OnLine 05/2015; 14(1). DOI:10.1186/s12938-015-0043-3 · 1.75 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The conventional approach to evaluate female pattern hair loss (FPHL) is to visually inspect and score images of balding area (BA). However, visual estimates vary widely among different physicians, and may hinder objective assessment of hair loss and subsequent treatment response. For this reason, we propose a quantitative method using a computer-aided imaging system to help physicians evaluate the severity of FPHL clinically.
    Dermatologica Sinica 02/2015; DOI:10.1016/j.dsi.2015.01.002 · 0.57 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Purpose. Most applications in the field of medical image processing require precise estimation. To improve the accuracy of segmentation, this study aimed to propose a novel segmentation method for coronary arteries to allow for the automatic and accurate detection of coronary pathologies. Methods. The proposed segmentation method included 2 parts. First, 3D region growing was applied to give the initial segmentation of coronary arteries. Next, the location of vessel information, HHH subband coefficients of the 3D DWT, was detected by the proposed vessel-texture discrimination algorithm. Based on the initial segmentation, 3D DWT integrated with the 3D neutrosophic transformation could accurately detect the coronary arteries. Results. Each subbranch of the segmented coronary arteries was segmented correctly by the proposed method. The obtained results are compared with those ground truth values obtained from the commercial software from GE Healthcare and the level-set method proposed by Yang et al., 2007. Results indicate that the proposed method is better in terms of efficiency analyzed. Conclusion. Based on the initial segmentation of coronary arteries obtained from 3D region growing, one-level 3D DWT and 3D neutrosophic transformation can be applied to detect coronary pathologies accurately.
    BioMed Research International 01/2015; 2015:798303. DOI:10.1155/2015/798303 · 2.71 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Recent studies suggest that structural and functional alterations of the language network are associated with auditory verbal hallucinations (AVHs) in schizophrenia. However, the ways in which the underlying structure and function of the network are altered and how these alterations are related to each other remain unclear. To elucidate this, we used diffusion spectrum imaging (DSI) to reconstruct the dorsal and ventral pathways and employed functional magnetic resonance imaging (fMRI) in a semantic task to obtain information about the functional activation in the corresponding regions in 18 patients with schizophrenia and 18 matched controls. The results demonstrated decreased structural integrity in the left ventral, right ventral and right dorsal tracts, and decreased functional lateralization of the dorsal pathway in schizophrenia. There was a positive correlation between the microstructural integrity of the right dorsal pathway and the functional lateralization of the dorsal pathway in patients with schizophrenia. Additionally, both functional lateralization of the dorsal pathway and microstructural integrity of the right dorsal pathway were negatively correlated with the scores of the delusion/hallucination symptom dimension. Our results suggest that impaired structural integrity of the right dorsal pathway is related to the reduction of functional lateralization of the dorsal pathway, and these alterations may aggravate AVHs in schizophrenia.
    Psychiatry Research Neuroimaging 09/2014; DOI:10.1016/j.pscychresns.2014.08.010 · 2.83 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: This work presents a new method for segmenting coronary arteries automatically in computed tomography angiography (CTA) data sets. The method automatically isolates heart and coronary arteries from surrounding structures and search for the probable location of coronary arteries by 3D region growing. Based on the dilation of the probable location, discrete wavelet transformation (DWT) and λ - mean operation complete accurate detection of coronary arties. Finally, the proposed method is tested on clinical CTA data-sets. The results on clinical datasets show that the proposed method is able to extract each branch of arteries when comparing to commercial software GE Healthcare and delineated ground truth.
    Journal of Medical Systems 06/2014; 38(6):55. DOI:10.1007/s10916-014-0055-8 · 2.21 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We present the first experimental results of time-resolved diffuser-aided diffuse optical imaging (DADOI) method in this paper. A self-manufactured diffuser plate was inserted between the optode and the surface of a scattering medium. The diffuser was utilized to enhance the multiple scattering that destroys the image information for baseline measurement of turbid medium. Therefore, the abnormality can be detected with the modified optical density calculation. The time-domain DADOI method can provide better imaging contrast and simpler imaging than the conventional diffuse optical tomography measurement. Besides, it also reveals rich depth information with temporal responses. Therefore, the DADOI offers a great potential to detect the breast tumor and chemotherapy monitoring in clinical diagnosis.
    Journal of Biomedical Optics 04/2014; 19(4):46008. DOI:10.1117/1.JBO.19.4.046008 · 2.75 Impact Factor
  • Chia-Yun Hsu, Yi-Hong Chou, Chung-Ming Chen
    [Show abstract] [Hide abstract]
    ABSTRACT: A whole breast ultrasound tumor detection algorithm, which could help physicians determine and diagnose, has been proposed. First of all, we employ the multi-scale blob detection. Secondly, we discard most unlikely lesions by ultrasound confidence maps and sheet detection, according to the prior knowledge of breast anatomy. After discarding unlikely blob structures, we collect the survival blob-like structures from the results of four passes of multi-scale blob detection in a single set. The features of blobness, size, and probability of being at muscle layer of the all detected blob-like candidates are used in a classification process for the differentiation of true lesions from negative ones in the collection set. Mutual information based feature selection (MIFS) procedure is applied to select three most effective features for the purpose of dimension reduction with the aid of logistic regression classifier and the process of leave-one-out cross-validation (LOO-CV) method. 49 datasets were acquired from 29 patients, which contain 86 lesions in total. According to the area under the ROC curve (AUC), the technique shows promising future for computer aided detection and the superior results when compared to Moon's method.
    2014 International Conference on Computational Science and Computational Intelligence (CSCI); 03/2014
  • Mathematical Problems in Engineering 01/2014; 2014:1-12. DOI:10.1155/2014/902659 · 1.08 Impact Factor
  • Computational and Mathematical Methods in Medicine 09/2013; 2013:790608. DOI:10.1155/2013/790608 · 1.02 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Perfusion computed tomography (CT) has been widely used to assess the response of lung cancer treatment. However, the respiratory motion has become the major obstacle to the pixel-based time-series analyses. To minimize the effect of respiratory motion and investigate the feasibility of perfusion CT for prediction of tumor response and prognosis of non-small cell lung cancer, an image registration framework is proposed by unifying a virtual 3D local rigid alignment and 3D global non-rigid alignment. The basic idea is to use the perfusion CT data and routine whole-lung CT data, respectively. To realize this idea, maximum intensity projection (MIP) of the time series perfusion CT images is first generated, followed by decomposing the MIP image into region of interest (ROI), which is located on a lung nodule. For the ROI, affine transformation model based on mutual information is performed to estimate the virtual three dimensional linear deformations. Following that, the 3D thin plate spline (TPS) is carried out to establish the pixel correspondence between the paired volumetric CT data. The control points for the TPS are global feature points chosen from the boundary of whole lung, which are automatically derived by using the iterative closest point (ICP) matching Algorithm. The proposed algorithm has been evaluated both qualitatively and quantitatively on real lung perfusion CT datasets. From the time-intensity curves and perfusion parameters, the experiment results suggest that the findings on perfusion CT images obtained after treatment may be considered as a significant predictor of lung cancer.
    SPIE Medical Imaging; 03/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: Image segmentation for the demarcation of pulmonary nodules in CT images is intrinsically an arduous task. The difficulty can be summarized into two aspects. Firstly, lung tumor can be various in terms of physical densities in pulmonary regions, implying the different interpretation as a mixture of GGO and solid nodules. Hence, processing of lung CT images may generally encounter tissue inhomogeneous problem. The second factor that complicates the task of nodule demarcation is the irregular shapes that most nodules are directly connected to other structures sharing the similar density profile. In this paper, an image segmentation framework is proposed by unifying the techniques of statistical region merging and conditional random field (CRF) with graph cut optimization to address the difficult problem of GGO nodules quantification in CT images. Different from traditional segmentation methods that use pixel-based approach such as region growing and morphological constraints, we employ a hierarchical segmentation tree to alleviate the effect of inhomogeneous attenuation. In addition to building perceptual prominent regions, we perform inference in CRF model based on restricting the pool of segmented regions. Following that, an inference CRF model is carried out to detect and localize individual object instances in CT images. The proposed algorithm is evaluated with four sets of manual delineations on 77 lung CT images. Incorporating with the efficiency and accuracy of pulmonary nodules segmentation method proposed in this paper, a computer aided system is hence feasible to develop related clinical application.
    SPIE Medical Imaging; 03/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: The goal of this study is to prove that the light propagation in the head by used the 3-D optical model from in vivo MRI data set can also provide significant characteristics on the spatial sensitivity of cerebral cortex folding geometry based on Monte Carlo simulation. Thus, we proposed a MRI based approach for 3-D brain modeling of near-infrared spectroscopy (NIRS). In the results, the spatial sensitivity profile of the cerebral cortex folding geometry and the arrangement of source-detector separation have being necessarily considered for applications of functional NIRS. The optimal choice of source-detector separation is suggested within 3-3.5 cm by the received intensity with different source-detector separations and the ratio of received light from the gray and white matter layer is greater than 50%. Additionally, this study has demonstrated the capability of NIRS in not only assessing the functional but also detecting the structural change of the brain by taking advantage of the low scattering and absorption coefficients observed in CSF of sagittal view. (© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim).
    Journal of Biophotonics 03/2013; 6(3). DOI:10.1002/jbio.201200025 · 3.86 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes a general boundary delineation method for 2D serial US images by modeling the spatial-temporal dynamics and generating the object boundary in each slice under the cell-based MAP scheme. The modeling of the spatial-temporal dynamics can serve as a prior to guide the cell-based MAP process and potentially maintain the contextual coherence. Experiments have been conducted on 8 sets of compression breast series and 5 sets of freehand breast acquisitions. The computer-generated results by our algorithm are compared to manual delineations prepared by experts. The experimental results suggest that the boundaries of the proposed method are not significantly different to manual outlines and are quite stable in terms of reproducibility.
    Image and Graphics (ICIG), 2013 Seventh International Conference on; 01/2013
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: BACKGROUND: The purpose of this study was to analyze the relationship between cytologic features, clinical features, and recurrence of papillary thyroid cancer (PTC). We hoped to predict prognosis preoperatively. METHODS: We retrospectively studied cytologic features by using computerized morphometry and clinical data of 118 patients with usual-type PTC without initial metastasis, including 34 patients with cancer recurrence in 10 years after surgery and 84 patients who did not have recurrence for more than 10 years after surgery. Another 24 patients were recruited for validation. RESULTS: Multivariate logistic analysis indicated that nucleus-to-cell ratio, variation of nuclear area, tumor size, and patient age were significantly related to recurrence. Cox regression analysis showed that hazard ratios were 3.34, 1.53, 1.77, and 2.6, respectively. CONCLUSION: Cytologic features of PTC analyzed with computerized morphometry significantly correlated with recurrence. It helped to predict prognosis preoperatively and may be helpful for planning further treatment. © 2012 Wiley Periodicals, Inc. Head Neck, 2012.
    Head & Neck 01/2013; 35(1). DOI:10.1002/hed.22909 · 3.01 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: To investigate dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) of advanced nonsmall-cell lung cancer (NSCLC) patients treated with the antiangiogenic agent bevacizumab combined with gemcitabine and cisplatin as first-line treatment. All patients were enrolled for MRI and computed tomography (CT) before and after the first three courses of bevacizumab combination chemotherapy. Pharmacokinetic parameters (K(trans), k(ep), v(e), v(p)) derived from DCE MRI were computed for the main mass. Parametric histogram analysis was obtained to evaluate changes of the internal tumor composition and for correlation with tumor response measured on CT. After three cycles of treatment, 11 patients showed decreased tumor size and a decreased value of all MR-derived pharmacokinetic parameters. Among these parameters, there was a significant decrease of mean and standard deviation of the K(trans) histogram as well as a decrease of mean of the k(ep) histogram (P < 0.05). Tumors with larger mean values of rate constant k(ep) (P < 0.0001) and smaller standard deviation of volume of extravascular extracellular space fraction v(e) (P < 0.0001) on histograms before chemotherapy were considered predictors for treatment response. DCE MRI enables a functional analysis of the treatment response of NSCLC. MRI parametric histogram has the potential to predict early treatment response of combined bevacizumab, gemcitabine, and cisplatin.
    Journal of Magnetic Resonance Imaging 08/2012; 36(2):387-96. DOI:10.1002/jmri.23660 · 2.79 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Identification of possible protein targets of small chemical molecules is an important step for unravelling their underlying causes of actions at the molecular level. To this end, we construct a web server, idTarget, which can predict possible binding targets of a small chemical molecule via a divide-and-conquer docking approach, in combination with our recently developed scoring functions based on robust regression analysis and quantum chemical charge models. Affinity profiles of the protein targets are used to provide the confidence levels of prediction. The divide-and-conquer docking approach uses adaptively constructed small overlapping grids to constrain the searching space, thereby achieving better docking efficiency. Unlike previous approaches that screen against a specific class of targets or a limited number of targets, idTarget screen against nearly all protein structures deposited in the Protein Data Bank (PDB). We show that idTarget is able to reproduce known off-targets of drugs or drug-like compounds, and the suggested new targets could be prioritized for further investigation. idTarget is freely available as a web-based server at http://idtarget.rcas.sinica.edu.tw.
    Nucleic Acids Research 05/2012; 40(Web Server issue):W393-9. DOI:10.1093/nar/gks496 · 9.11 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Genetic/transcriptional regulatory interactions are shown to predict partial components of signaling pathways, which have been recognized as vital to complex human diseases. Both activator (A) and repressor (R) are known to coregulate their common target gene (T). Xu et al. (2002) proposed to model this coregulation by a fixed second order response surface (called the RS algorithm), in which T is a function of A, R, and AR. Unfortunately, the RS algorithm did not result in a sufficient number of genetic interactions (GIs) when it was applied to a group of 51 yeast genes in a pilot study. Thus, we propose a data-driven second order model (DDSOM), an approximation to the non-linear transcriptional interactions, to infer genetic and transcriptional regulatory interactions. For each triplet of genes of interest (A, R, and T), we regress the expression of T at time t + 1 on the expression of A, R, and AR at time t. Next, these well-fitted regression models (viewed as points in R(3)) are collected, and the center of these points is used to identify triples of genes having the A-R-T relationship or GIs. The DDSOM and RS algorithms are first compared on inferring transcriptional compensation interactions of a group of yeast genes in DNA synthesis and DNA repair using microarray gene expression data; the DDSOM algorithm results in higher modified true positive rate (about 75%) than that of the RS algorithm, checked against quantitative RT-polymerase chain reaction results. These validated GIs are reported, among which some coincide with certain interactions in DNA repair and genome instability pathways in yeast. This suggests that the DDSOM algorithm has potential to predict pathway components. Further, both algorithms are applied to predict transcriptional regulatory interactions of 63 yeast genes. Checked against the known transcriptional regulatory interactions queried from TRANSFAC, the proposed also performs better than the RS algorithm.
    Frontiers in Genetics 05/2012; 3:71. DOI:10.3389/fgene.2012.00071
  • [Show abstract] [Hide abstract]
    ABSTRACT: The symptom of brain volumetric changes may provide significant biomarker to predict progressive dementia. The brain volumetric changes of prefrontal cortex are highly associated with many neurodegenerative diseases. Besides, brain atrophy reveals the expanded interhemispheric fissure and the concomitant increasing cerebrospinal fluid volume. Thus, the quantitative assessment of brain volumetric changes is an important consideration for clinical studies of neurodegenerative diseases. In this study, we first proposed an approach that uses near-infrared brain volumetric imaging to detect brain volumetric changes. The healthy, aged, and typical Alzheimer's disease (AD) brains were modeled with different characterization of brain volumetric changes from in vivo MRI data based on time-resolved 3-D Monte Carlo simulation. In the results, the significant difference of prefrontal cortex structure can be observed among healthy, aged, and AD brain with various source-detector separations in sagittal view. Our study shows that the near-infrared brain volumetric imaging can be an indicator of brain atrophy for clinical application of neurodegenerative diseases with patient-oriented measurement.
    IEEE Journal of Selected Topics in Quantum Electronics 05/2012; 18(3):1122-1129. DOI:10.1109/JSTQE.2011.2163927 · 3.47 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A probabilistic image segmentation algorithm called stochastic region competition is proposed for performing Doppler sonography segmentation. The image segmentation is conducted by maximizing a posteriori that models histogram likelihood, gradient likelihood, and a spatial prior. The optimization is done by a modified expectation and maximization (EM) method that aims to improve computation efficiency and avoid local optima. The algorithm was tested on 155 color Doppler sonograms and compared with manual delineations. The qualitative assessment shows that our algorithm is able to segment mass lesions under the condition of low image quality and the interference of the color-encoded Doppler information. The quantitative assessment analysis shows that the average distance between the algorithm-generated boundaries and manual delineations is statistically comparable to the variability of manual delineations. The ratio of the overlapping area between the algorithm-generated boundaries and manual delineations is also comparable to that between different sets of manual delineations. A reproductivity test was conducted to confirm that the result is statistically reproducible. The algorithm can be used to perform Doppler sonography segmentation and to replace the tedious manual delineation task in clinical application.
    Medical Physics 05/2012; 39(5):2867-76. DOI:10.1118/1.4705350 · 3.01 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: This study proposed diffuser-aided diffuse optical imaging (DADOI) as a new approach to improve the performance of the conventional diffuse optical tomography (DOT) approach for breast imaging. The 3-D breast model for Monte Carlo simulation is remodeled from clinical MRI image. The modified Beer-Lambert's law is adopted with the DADOI approach to substitute the complex algorithms of inverse problem for mapping of spatial distribution, and the depth information is obtained based on the time-of-flight estimation. The simulation results demonstrate that the time-resolved Monte Carlo method can be capable of performing source-detector separations analysis. The dynamics of photon migration with various source-detector separations are analyzed for the characterization of breast tissue and estimation of optode arrangement. The source-detector separations should be less than 4 cm for breast imaging in DOT system. Meanwhile, the feasibility of DADOI was manifested in this study. In the results, DADOI approach can provide better imaging contrast and faster imaging than conventional DOT measurement. The DADOI approach possesses great potential to detect the breast tumor in early stage and chemotherapy monitoring that implies a good feasibility for clinical application.
    IEEE transactions on bio-medical engineering 05/2012; 59(5):1454-61. DOI:10.1109/TBME.2012.2187900 · 2.23 Impact Factor

Publication Stats

342 Citations
100.99 Total Impact Points

Institutions

  • 2003–2015
    • National Taiwan University
      • Institute of Biomedical Engineering
      T’ai-pei, Taipei, Taiwan
  • 2012
    • Carnegie Mellon University
      • Department of Biomedical Engineering
      Pittsburgh, PA, United States
  • 2008
    • Chaoyang University of Technology
      臺中市, Taiwan, Taiwan