Volume 1 Issue 3
Authors: Uygar Tuna; Ulla Ruotsalainen; Antti Sohlberg
Abstract: The aim of this study was to investigate: could the number of projection angles, and thus the total acquisition time be reduced using maximum a posteriori expectation maximization (MAP-EM) method without loss in the quality of the reconstructed images? In this study, we evaluate the sequentially applied MAP-EM method in which the amount of regularization is reduced step-by-step during the reconstruction process. The method was used for the reconstruction of the SPECT data with few projection angles and at low count levels. We assessed the MAP-EM method with three different spatial domain regularizers: (1) Median filter, (2) L-filter and (3) Block-Matching and 3D filter. The performance of the MAP-EM method was examined with numerical cardiac phantom and real patient data. We used a broad test dataset with different amount of projection angles at different count levels. The volumes reconstructed with the MAP-EM (with different regularizers) were compared to the images reconstructed with the maximum likelihood expectation maximization (MLEM) method. In addition to the visual evaluations, we gave comparisons using the contrast ratio and coefficient of variation measures. We also examined the reconstructed images via profiles drawn through the cardiac region. The images reconstructed with the sequentially applied MAP-EM method were visually very good and showed hardly any visual differences for most of the cases. The quantitative evaluations showed that for all cases, the reconstructed images were improved with the MAP-EM reconstruction method compared to the MLEM. For certain range of counts depending on the study and scanner properties, collecting the counts into fewer projection angles posed advantage on collecting the same amount of counts into more projection angles. With the inclusion of the sequential application of the regularization filter in the MAP-EM method, it is possible to reduce SPECT acquisition time down to certain levels depending on the collected amount of counts and numbers of projection angles.
Keywords: Analytical Reconstruction; Block Matching and 3D Filtering; Bone Scan; Contrast Ratio; Few Angle Projection; Filtered Backprojection; Ill-Posed; Incomplete Sinogram; Inverse Problem; Maximum Likelihood Expectation Maximization; Median Filtering
Authors: Behnaz Ghoraani
Abstract: While time-frequency (TF) representations are commonly used for visualization purposes, the adaptive feature extraction from TF matrices has not been extensively studied in the literature. Even when used, there exists no unique or automated methodology to extract discriminative TF features from non-stationary signals. The present paper focuses on feature extraction from time-frequency distribution (TFD), and attempts to develop a generalized TF matrix (TFM) analysis methodology that exploits the benefits of TFD in pattern classification systems. The proposed TFM feature extraction consists of two stages: first, we build TFM decomposition that uses a matrix decomposition (MD) technique to effectively segment non-stationary signals. Second, instantaneous and unique features are extracted from each segment in a way that they successfully represent joint TF structure of the signal. In this paper, the suitable tools for TFM feature extraction pertaining to pattern classification are investigated through experiments with different synthetic signals. Experiments performed with pathological speech and environmental audio signals produced 98.6% and 85.5% accuracy rates respectively demonstrating the benefits of TFM feature extraction for pattern classification related applications.
Keywords: Time-Frequency Matrix Analysis; Time-Frequency Analysis; Time-Frequency Feature Extraction; Matrix Decomposition
Authors: Nouhoun Kane; Ahmed El Oirrak
Abstract: In this paper we use Discrete Wavelet Decomposition (DWT) for comparing document. First documents are transformed to ASCII code; then for each level DTW coefficients are extracted from ASCII code. Performances are measured using real’s documents. A simple comparison is made between DWT and Bag of Word (BOW) representation to show goodness of proposed technique.
Keywords: Component; Text Representation; Discrete Wavelet Transform (DWT); Bag of Word( BOW) Representation