Sema Candemir
National Institutes of Health
Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Sema Candemir.
IEEE Transactions on Medical Imaging | 2014
Sema Candemir; Stefan Jaeger; Kannappan Palaniappan; Jonathan P. Musco; Rahul Singh; Zhiyun Xue; Alexandros Karargyris; Sameer K. Antani; George R. Thoma; Clement J. McDonald
The National Library of Medicine (NLM) is developing a digital chest X-ray (CXR) screening system for deployment in resource constrained communities and developing countries worldwide with a focus on early detection of tuberculosis. A critical component in the computer-aided diagnosis of digital CXRs is the automatic detection of the lung regions. In this paper, we present a nonrigid registration-driven robust lung segmentation method using image retrieval-based patient specific adaptive lung models that detects lung boundaries, surpassing state-of-the-art performance. The method consists of three main stages: 1) a content-based image retrieval approach for identifying training images (with masks) most similar to the patient CXR using a partial Radon transform and Bhattacharyya shape similarity measure, 2) creating the initial patient-specific anatomical model of lung shape using SIFT-flow for deformable registration of training masks to the patient CXR, and 3) extracting refined lung boundaries using a graph cuts optimization approach with a customized energy function. Our average accuracy of 95.4% on the public JSRT database is the highest among published results. A similar degree of accuracy of 94.1% and 91.7% on two new CXR datasets from Montgomery County, MD, USA, and India, respectively, demonstrates the robustness of our lung segmentation approach.
IEEE Transactions on Medical Imaging | 2014
Stefan Jaeger; Alexandros Karargyris; Sema Candemir; Les R. Folio; Jenifer Siegelman; Fiona M. Callaghan; Zhiyun Xue; Kannappan Palaniappan; Rahul K. Singh; Sameer K. Antani; George R. Thoma; Yi-Xiang J. Wang; Pu-Xuan Lu; Clement J. McDonald
Tuberculosis is a major health threat in many regions of the world. Opportunistic infections in immunocompromised HIV/AIDS patients and multi-drug-resistant bacterial strains have exacerbated the problem, while diagnosing tuberculosis still remains a challenge. When left undiagnosed and thus untreated, mortality rates of patients with tuberculosis are high. Standard diagnostics still rely on methods developed in the last century. They are slow and often unreliable. In an effort to reduce the burden of the disease, this paper presents our automated approach for detecting tuberculosis in conventional posteroanterior chest radiographs. We first extract the lung region using a graph cut segmentation method. For this lung region, we compute a set of texture and shape features, which enable the X-rays to be classified as normal or abnormal using a binary classifier. We measure the performance of our system on two datasets: a set collected by the tuberculosis control program of our local countys health department in the United States, and a set collected by Shenzhen Hospital, China. The proposed computer-aided diagnostic system for TB screening, which is ready for field deployment, achieves a performance that approaches the performance of human experts. We achieve an area under the ROC curve (AUC) of 87% (78.3% accuracy) for the first set, and an AUC of 90% (84% accuracy) for the second set. For the first set, we compare our system performance with the performance of radiologists. When trying not to miss any positive cases, radiologists achieve an accuracy of about 82% on this set, and their false positive rate is about half of our systems rate.
Quantitative imaging in medicine and surgery | 2013
Stefan Jaeger; Alexandros Karargyris; Sema Candemir; Jenifer Siegelman; Les R. Folio; Sameer K. Antani; George R. Thoma
Tuberculosis (TB) is a major global health threat. An estimated one-third of the worlds population has a history of TB infection, and millions of new infections are occurring every year. The advent of new powerful hardware and software techniques has triggered attempts to develop computer-aided diagnostic systems for TB detection in support of inexpensive mass screening in developing countries. In this paper, we describe the medical background of TB detection in chest X-rays and present a survey of the recent approaches using computer-aided detection. After a thorough research of the computer science literature for such systems or related methods, we were able to identify 16 papers, including our own, written between 1996 and early 2013. These papers show that TB screening is a challenging task and an open research problem. We report on the progress to date and describe experimental screening systems that have been developed.
Quantitative imaging in medicine and surgery | 2014
Stefan Jaeger; Sema Candemir; Sameer K. Antani; Yi-Xiang J. Wang; Pu-Xuan Lu; George R. Thoma
The U.S. National Library of Medicine has made two datasets of postero-anterior (PA) chest radiographs available to foster research in computer-aided diagnosis of pulmonary diseases with a special focus on pulmonary tuberculosis (TB). The radiographs were acquired from the Department of Health and Human Services, Montgomery County, Maryland, USA and Shenzhen No. 3 Peoples Hospital in China. Both datasets contain normal and abnormal chest X-rays with manifestations of TB and include associated radiologist readings.
international conference on image analysis and recognition | 2010
Sema Candemir; Yusuf Sinan Akgul
Graph cut minimization formulates the segmentation problem as the liner combination of data and smoothness terms. The smoothness term is included in the energy formulation through a regularization parameter. We propose that the trade-off between the data and the smoothness terms should not be balanced by the same regularization parameter for the whole image. In order to validate the proposed idea, we build a system which adaptively changes the effect of the regularization parameter for the graph cut segmentation. The method calculates the probability of being part of the boundary for each pixel using the Canny edge detector at different hysteresis threshold levels. Then, it adjusts the regularization parameter of the pixel depending on the probability value. The experiments showed that adjusting the effect of the regularization parameter on different image regions produces better segmentation results than using a single best regularization parameter.
international symposium on biomedical imaging | 2013
Sema Candemir; Kannappan Palaniappan; Yusuf Sinan Akgul
One of the first steps of computer-aided systems is robustly detect the anatomical boundaries. Literature has several successful energy minimization based algorithms which are applied to medical images. However, these algorithms depend on parameters which need to be tuned for a meaningful solution. One of the important parameters is the regularization parameter (λ) which is generally estimated in an ad-hoc manner and is used for the whole data set. In this paper we claim that λ can be learned by local features which hold the regional characteristics of the image. We propose a λ estimation system which is modeled as a multi-class classification scheme. We demonstrate the performance of the approach within graph cut segmentation framework via qualitative results on chest X-rays. Experimental results indicate that predicted parameters produce better segmentation results.
computer-based medical systems | 2015
Zhiyun Xue; Daekeun You; Sema Candemir; Stefan Jaeger; Sameer K. Antani; L. Rodney Long; George R. Thoma
The view information of a chest X-ray (CXR), such as frontal or lateral, is valuable in computer aided diagnosis (CAD) of CXRs. For example, it helps for the selection of atlas models for automatic lung segmentation. However, very often, the image header does not provide such information. In this paper, we present a new method for classifying a CXR into two categories: frontal view vs. lateral view. The method consists of three major components: image pre-processing, feature extraction, and classification. The features we selected are image profile, body size ratio, pyramid of histograms of orientation gradients, and our newly developed contour-based shape descriptor. The method was tested on a large (more than 8,200 images) CXR dataset hosted by the National Library of Medicine. The very high classification accuracy (over 99% for 10-fold cross validation) demonstrates the effectiveness of the proposed method.
Proceedings of SPIE | 2012
Sema Candemir; Kannappan Palaniappan; Filiz Bunyak
Aerial wide-area monitoring and tracking using multi-camera arrays poses unique challenges compared to stan- dard full motion video analysis due to low frame rate sampling, accurate registration due to platform motion, low resolution targets, limited image contrast, static and dynamic parallax occlusions.1{3 We have developed a low frame rate tracking system that fuses a rich set of intensity, texture and shape features, which enables adaptation of the tracker to dynamic environment changes and target appearance variabilities. However, improper fusion and overweighting of low quality features can adversely aect target localization and reduce tracking performance. Moreover, the large computational cost associated with extracting a large number of image-based feature sets will in uence tradeos for real-time and on-board tracking. This paper presents a framework for dynamic online ranking-based feature evaluation and fusion in aerial wide-area tracking. We describe a set of ecient descriptors suitable for small sized targets in aerial video based on intensity, texture, and shape feature representations or views. Feature ranking is then used as a selection procedure where target-background discrimination power for each (raw) feature view is scored using a two-class variance ratio approach. A subset of the k-best discriminative features are selected for further processing and fusion. The target match probability or likelihood maps for each of the k features are estimated by comparing target descriptors within a search region using a sliding win- dow approach. The resulting k likelihood maps are fused for target localization using the normalized variance ratio weights. We quantitatively measure the performance of the proposed system using ground-truth tracks within the framework of our tracking evaluation test-bed that incorporates various performance metrics. The proposed feature ranking and fusion approach increases localization accuracy by reducing multimodal eects due to low quality features or background clutter. Adaptive feature ranking increases the robustness of the tracker in dynamically changing environments especially when the object appearance is changing.
International Journal of Pattern Recognition and Artificial Intelligence | 2015
K. C. Santosh; Sema Candemir; Stefan Jaeger; Alexandros Karargyris; Sameer K. Antani; George R. Thoma; Les R. Folio
We present a novel method for detecting rotated lungs in chest radiographs for quality control and augmenting automated abnormality detection. The method computes a principal rib-orientation measure using a generalized line histogram technique for quality control, and therefore augmenting automated abnormality detection. To compute the line histogram, we use line seed filters as kernels to convolve with edge images, and extract a set of lines from the posterior rib-cage. After convolving kernels in all possible orientations in the range [0°, 180°), we measure the angle with maximum magnitude in the line histogram. This measure provides an approximation of the principal chest rib-orientation for each lung. A chest radiograph is upright if the difference between the orientation angles of both lungs with respect to the horizontal axis is negligible. We validate our method on sets of normal and abnormal images and argue that rib orientation can be used for rotation detection in chest radiographs as an aid in quality control during image acquisition. It can also be used for training and testing data sets for computer aided diagnosis research, for example. In our experiments, we achieve a maximum accuracy of approximately 90%.
Image and Vision Computing | 2015
Sema Candemir; Eugene Borovikov; K. C. Santosh; Sameer K. Antani; George R. Thoma
Abstract Modern appearance-based object recognition systems typically involve feature/descriptor extraction and matching stages. The extracted descriptors are expected to be robust to illumination changes and to reasonable (rigid or affine) image/object transformations. Some descriptors work well for general object matching, but gray-scale key-point-based methods may be suboptimal for matching line-rich color scenes/objects such as cars, buildings, and faces. We present a Rotation- and Scale-Invariant, Line-based Color-aware descriptor (RSILC), which allows matching of objects/scenes in terms of their key-lines, line-region properties, and line spatial arrangements. An important special application is face matching, since face characteristics are best captured by lines/curves. We tested RSILC performance on publicly available datasets and compared it with other descriptors. Our experiments show that RSILC is more accurate in line-rich object description than other well-known generic object descriptors.