Current Issue : April - June Volume : 2012 Issue Number : 2 Articles : 5 Articles
The registration of intraoperative ultrasound (US) images with preoperative magnetic resonance (MR) images is a challenging\r\nproblem due to the difference of information contained in each image modality. To overcome this difficulty, we introduce a new\r\nprobabilistic function based on the matching of cerebral hyperechogenic structures. In brain imaging, these structures are the\r\nliquid interfaces such as the cerebral falx and the sulci, and the lesions when the corresponding tissue is hyperechogenic. The\r\nregistration procedure is achieved by maximizing the joint probability for a voxel to be included in hyperechogenic structures in\r\nboth modalities. Experiments were carried out on real datasets acquired during neurosurgical procedures. The proposed validation\r\nframework is based on (i) visual assessment, (ii) manual expert estimations , and (iii) a robustness study. Results show that\r\nthe proposed method (i) is visually efficient, (ii) produces no statistically different registration accuracy compared to manualbased\r\nexpert registration, and (iii) converges robustly. Finally, the computation time required by our method is compatible with\r\nintraoperative use....
Label fusion is used in medical image segmentation to combine several different labels of the same entity into a single discrete\r\nlabel, potentially more accurate, with respect to the exact, sought segmentation, than the best input element. Using simulated data,\r\nwe compared three existing label fusion techniquesââ?¬â?STAPLE, Voting, and Shape-Based Averaging (SBA)ââ?¬â?and observed that\r\nnone could be considered superior depending on the dissimilarity between the input elements. We thus developed an empirical,\r\nhybrid technique called SVS, which selects the most appropriate technique to apply based on this dissimilarity. We evaluated the\r\nlabel fusion strategies on two- and three-dimensional simulated data and showed that SVS is superior to any of the three existing\r\nmethods examined. On real data, we used SVS to perform fusions of 10 segmentations of the hippocampus and amygdala in 78\r\nsubjects from the ICBM dataset. SVS selected SBA in almost all cases, which was the most appropriate method overall....
Attention is crucial for encoding information into memory, and current dual-process models seek to explain the roles of attention\nin both recollection memory and incidental-perceptual memory processes. The present study combined an incidental memory\nparadigm with event-related functional MRI to examine the effect of attention at encoding on the subsequent neural activation\nassociated with unintended perceptual memory for spoken words. At encoding, we systematically varied attention levels as listeners\nheard a list of single English nouns.We then presented these words again in the context of a recognition task and assessed the effect\nof modulating attention at encoding on the BOLD responses to words that were either attended strongly, weakly, or not heard\npreviously. MRI revealed activity in right-lateralized inferior parietal and prefrontal regions, and positive BOLD signals varied\nwith the relative level of attention present at encoding. Temporal analysis of hemodynamic responses further showed that the time\ncourse of BOLD activity was modulated differentially by unintentionally encoded words compared to novel items. Our findings\nlargely support current models of memory consolidation and retrieval, but they also provide fresh evidence for hemispheric\ndifferences and functional subdivisions in right frontoparietal attention networks that help shape auditory episodic recall....
Machine learning (ML) plays an important role in the medical imaging field, including medical image analysis and computeraided\r\ndiagnosis, because objects such as lesions and organs may not be represented accurately by a simple equation; thus, medical\r\npattern recognition essentially require ââ?¬Å?learning from examples.ââ?¬Â One of the most popular uses of ML is classification of objects\r\nsuch as lesions into certain classes (e.g., abnormal or normal, or lesions or nonlesions) based on input features (e.g., contrast\r\nand circularity) obtained from segmented object candidates. Recently, pixel/voxel-based ML (PML) emerged in medical image\r\nprocessing/analysis, which use pixel/voxel values in images directly instead of features calculated from segmented objects as input\r\ninformation; thus, feature calculation or segmentation is not required. Because the PML can avoid errors caused by inaccurate\r\nfeature calculation and segmentation which often occur for subtle or complex objects, the performance of the PML can potentially\r\nbe higher for such objects than that of common classifiers (i.e., feature-based MLs). In this paper, PMLs are surveyed to make clear\r\n(a) classes of PMLs, (b) similarities and differences within (among) different PMLs and those between PMLs and feature-based\r\nMLs, (c) advantages and limitations of PMLs, and (d) their applications in medical imaging....
Consideration of information from multiple modalities has been shown to have increased diagnostic power in breast imaging.\r\nAs a result, new techniques such as microwave imaging continue to be developed. Interpreting these novel image modalities is\r\na challenge, requiring comparison to established techniques such as the gold standard X-ray mammography. However, due to\r\nthe highly deformable nature of breast tissues, comparison of 3D and 2D modalities is a challenge. To enable this comparison, a\r\nregistration technique was developed to map features from 2D mammograms to locations in the 3D image space. This technique\r\nwas developed and tested using magnetic resonance (MR) images as a reference 3Dmodality, asMRbreast imaging is an established\r\ntechnique in clinical practice. The algorithm was validated using a numerical phantom then successfully tested on twenty-four\r\nimage pairs. Dice�s coefficient was used to measure the external goodness of fit, resulting in an excellent overall average of 0.94.\r\nInternal agreement was evaluated by examining internal features in consultation with a radiologist, and subjective assessment\r\nconcludes that reasonable alignment was achieved....
Loading....