The Experts below are selected from a list of 198 Experts worldwide ranked by ideXlab platform
Hakan Erdogan - One of the best experts on this subject based on the ideXlab platform.
-
compressed sensing signal recovery via forward Backward pursuit
Digital Signal Processing, 2013Co-Authors: Nazim Burak Karahanoglu, Hakan ErdoganAbstract:Recovery of sparse signals from compressed measurements constitutes an @?"0 norm minimization problem, which is unpractical to solve. A number of sparse recovery approaches have appeared in the literature, including @?"1 minimization techniques, greedy pursuit algorithms, Bayesian methods and nonconvex optimization techniques among others. This manuscript introduces a novel two stage greedy approach, called the Forward-Backward Pursuit (FBP). FBP is an iterative approach where each iteration consists of consecutive forward and Backward stages. The forward Step first expands the support estimate by the forward Step size, while the following Backward Step shrinks it by the Backward Step size. The forward Step size is larger than the Backward Step size, hence the initially empty support estimate is expanded at the end of each iteration. Forward and Backward Steps are iterated until the residual power of the observation vector falls below a threshold. This structure of FBP does not necessitate the sparsity level to be known a priori in contrast to the Subspace Pursuit or Compressive Sampling Matching Pursuit algorithms. FBP recovery performance is demonstrated via simulations including recovery of random sparse signals with different nonzero coefficient distributions in noisy and noise-free scenarios in addition to the recovery of a sparse image.
-
forward Backward search for compressed sensing signal recovery
European Signal Processing Conference, 2012Co-Authors: Nazim Burak Karahanoglu, Hakan ErdoganAbstract:Reconstruction of sparse signals from reduced dimensions requires the solution of an l 0 norm minimization, which is unpractical. A number of algorithms have appeared in literature, including l 1 minimization, greedy pursuit algorithms, Bayesian methods and nonconvex optimization. This manuscript introduces a greedy approach, called the Forward-Backward Pursuit (FBP), which iteratively enlarges the support by consecutive forward and Backward Steps. At each iteration, the forward Step first expands the support, while the following Backward Step prunes it. The number of atoms selected by the forward Step is selected higher than the number of removals, hence the support is expanded at the end of each iteration. The recovery performance of the proposed method is demonstrated via simulations including different nonzero coefficient distributions in noisy and noise-free scenarios.
Vecchia, Claudio Dalla - One of the best experts on this subject based on the ideXlab platform.
-
Higher Order Hamiltonian Monte Carlo Sampling for Cosmological Large-Scale Structure Analysis
2021Co-Authors: Hernández-sánchez Mónica, Kitaura Francisco-shu, Ata Metin, Vecchia, Claudio DallaAbstract:We investigate higher order symplectic integration strategies within Bayesian cosmic density field reconstruction methods. In particular, we study the fourth-order discretisation of Hamiltonian equations of motion (EoM). This is achieved by recursively applying the basic second-order leap-frog scheme (considering the single evaluation of the EoM) in a combination of even numbers of forward time integration Steps with a single intermediate Backward Step. This largely reduces the number of evaluations and random gradient computations, as required in the usual second-order case for high-dimensional cases. We restrict this study to the lognormal-Poisson model, applied to a full volume halo catalogue in real space on a cubical mesh of 1250 $h^{-1}$ Mpc side and 256$^3$ cells. Hence, we neglect selection effects, redshift space distortions, and displacements. We note that those observational and cosmic evolution effects can be accounted for in subsequent Gibbs-sampling Steps within the COSMIC BIRTH algorithm. We find that going from the usual second to fourth-order in the leap-frog scheme shortens the burn-in phase by a factor of at least $\sim30$. This implies that 75-90 independent samples are obtained while the fastest second-order method converges. After convergence, the correlation lengths indicate an improvement factor of about 3.0 fewer gradient computations for meshes of 256$^3$ cells. In the considered cosmological scenario, the traditional leap-frog scheme turns out to outperform higher order integration schemes only at lower dimensional problems, e.g. meshes with 64$^3$ cells. This gain in computational efficiency can help to go towards a full Bayesian analysis of the cosmological large-scale structure for upcoming galaxy surveys.Comment: 19 pages, 12 figures, 4 tables, accepted at MNRAS, additional robust mathematical argument supported by numerical tests with longer HMC chains and a solid statistical analysi
-
Higher Order Hamiltonian Monte Carlo Sampling for Cosmological Large-Scale Structure Analysis
2020Co-Authors: Hernández-sánchez Mónica, Kitaura Francisco-shu, Ata Metin, Vecchia, Claudio DallaAbstract:We investigate higher order symplectic integration strategies within Bayesian cosmic density field reconstruction methods. In particular, we study the fourth-order discretisation of Hamiltonian equations of motion (EoM). This is achieved by recursively applying the basic second-order leap-frog scheme (considering the single evaluation of the EoM) in a combination of even numbers of forward time integration Steps with a single intermediate Backward Step. This largely reduces the number of evaluations and random gradient computations, as required in the usual second-order case. We restrict this study to the lognormal-Poisson model, applied to a full volume halo catalogue in real space on a cubical mesh of 1250 $h^{-1}$ Mpc side and 256$^3$ cells. Hence, we neglect selection effects, redshift space distortions, and displacements. We note that those observational and cosmic evolution effects can be accounted for in subsequent Gibbs-sampling Steps within the COSMIC BIRTH algorithm. We find that going from the usual second to fourth-order in the leap-frog scheme shortens the burn-in phase by a factor of at least $\sim30$. This implies that 80-100 independent samples are obtained while the fastest second-order method converges. After convergence, the correlation lengths indicate an improvement factor of $\sim2.4$ fewer gradient computations. This gain in computational efficiency can help to go towards a full Bayesian analysis of the cosmological large-scale structure for upcoming galaxy surveys.Comment: 17 pages, 10 figures, 4 table
Flávio Pulzatto - One of the best experts on this subject based on the ideXlab platform.
-
Atividade elétrica dos músculos estabilizadores da patela em indivíduos portadores da síndrome da dor femoropatelar durante exercícios realizados no Step.
Universidade Federal de São Carlos, 2005Co-Authors: Flávio PulzattoAbstract:A proposta deste estudo foi avaliar, nos aspectos temporal e de amplitude, a atividade elétrica dos músculos vasto medial oblíquo (VMO), vasto lateral longo (VLL) e vasto lateral oblíquo (VLO) nos exercícios de Step frontal: subida (SFS) e descida (SFD) e Step posterior: subida (SPS) e descida (SPD). Foram avaliados 27 indivíduos do sexo feminino divididos em dois grupos: 15 clinicamente normais Grupo Controle (21,13 2,17 anos) e 12 portadores da Síndrome da Dor Femoropatelar - SDFP (21,08 2,31). A altura do Step foi regulada para dois ângulos - 45 e 75 - de flexão da articulação do joelho. Um metrônomo auxiliou os voluntários quanto ao tempo de execução do exercício, um eletrogoniômetro foi utilizado para controlar o ângulo de flexão do joelho e um sensor de pressão foi utilizado para informar quanto ao início e o final de cada exercício. A atividade elétrica foi captada por meio de eletrodos ativos diferenciais simples de superfície, um eletromiógrafo de 8 canais (EMG System do Brasil) e um programa de aquisição de dados (AqDados 7.02.06). O sinal elétrico captado foi tratado por rotinas do software Matlab 6.1 que calcularam o tempo de início da ativação elétrica para cada músculo (análise temporal) e a integral matemática da área abaixo da envoltória do sinal retificado e filtrado (análise de amplitude). Os valores da integral foram normalizados pela média das três contrações para cada músculo e posteriormente calculada as relações VMO: VLO e VMO: VLL. O tempo relativo de ativação foi determinado subtraindo-se o tempo de ativação do VLL e do VLO do tempo de ativação do VMO (VMO - VLO e VMO-VLL). O teste t - Student (p ≤ 0,05) revelou que, no exercício de subida no Step a 75, houve diferença significativa no tempo relativo de ativação entre os grupos, tanto para o Step frontal: VMO-VLO (p = 0,000) e VMO-VLL (p = 0,000), quanto para o Step posterior: VMO-VLO (p = 0,000) e VMO-VLL (p = 0,000). No grupo Controle prevaleceu uma ativação antecipada do VMO em relação aos músculos VLL e VLO; já no grupo SDFP houve prevalência da ativação simultânea e tardia do VMO em relação aos demais músculos. A ANOVA thre-way e o teste de Duncan (p≤ 0,05) revelaram diferenças na relação VMO:VLO e VMO:VLL quando comparados os grupos Controle e SDFP (p = 0,014). Os valores da relação VMO:VLO e VMO:VLL foram significativamente maiores no Step a 45 do que a 75 (p = 0,000 e p = 0,016, respectivamente) nos dois grupos. A comparação entre os exercícios realizados dentro de uma mesma angulação de Step revelou que tanto a relação VMO:VLO quanto VMO:VLL sempre foram maiores no SFS quando comparado ao SFD (p = 0,010), ao SPS (p = 0,040) e ao SPD (p = 0,000). Não houve diferença entre a SFD e a SPS (p=0,570) ou a SPD (p = 0,090). No grupo Controle o SPS foi menor que a SPD (p = 0,030), enquanto que no grupo SDFP ocorreu o inverso nas duas relações em ambos os Steps. Nossos resultados sugerem haver diferenças no controle motor entre os grupos quanto ao recrutamento muscular, tanto no Step frontal quanto no posterior no ângulo de 75. Na análise da amplitude, o Step a 45 parece recrutar mais seletivamente o músculo VMO em relação ao VLL e VLO do que no Step a 75, podendo ser utilizado preferencialmente no tratamento de indivíduos portadores de SDFP. Quanto à modalidade de Step, o exercício de subida frontal (SFS) parece ser o mais indicado quando o objetivo for a ativação seletiva do músculo VMO principalmente no Step a 45.The purpose of this study was to evaluate the electric activity (EMG) in the temporal and amplitude aspects of the vastus medialis obliquus (VMO), vastus lateralis longus (VLL) and vastus lateralis obliquus (VLO) muscles during forward Step tasks: Stepdown (SFD) and Step-up (SFS), and Backward Step tasks: Step-down (SPD) and Step-up (SPS). Twenty seven females were evaluated and separated in two groups: fifteen normal subjects Control Group (21.13 2.17 years) and twelve subjects with Patellofemoral Pain Syndrome PPS (21.08 2.31 years). The height of the Step was regulated for two angles 45 and 75 - of knees flexion joint. A metronome was used to help the volunteers about the time to performance the tasks, a eletrogoniometer was used in the control of the knee angle and a pressure sensor was used to determinate the start and the end of the electromyography register. The electric activity was recorded by surface (Ag/AgCl) electrodes, an EMG apparatus with 8 channels (EMG System Brazil) and a software of acquisition data AqDados 7.02. The EMG data was processed by the software Matlab 6.1 that calculated both the onset timing of the muscles and the integrated of the EMG signal. The EMG was normalized by the mean of the three muscle contractions and was calculated de ratio VMO/VLL and VMO/VLO for comparisons between groups and exercises. The relative onset timing was determinate by the difference VMO-VLO and VMO-VLL. The t-tests showed that in the Step up exercise (75 of the knee flexion), there are differences when compared the groups, either in frontal Step VMO/VLL (p= 0.000) and VMO/VLO (p = 0.000), than posterior Step VMO/VLL (p = 0.000) and VMO/VLO (p = 0.000). In the control group there is a prevalence of an early contractions of de VMO muscle, however in the PPS group, the VMO onset occurred at the same time or after the VLO and VLL muscles in the mayor of the cases. The Anova three-way and Duncan post hoc showed that in Step at 45, the ratio VMO/VLO (p = 0,000) and VMO/VLL (p = 0,016) was greater then Step at 75 in both groups. When compared the Steps in 45 and 75 into the groups, were found differences either in the VMO:VLO ratio (p = 0,000) than VMO:VLL ratio (0,016) with greater values in the Step at 45. The comparison between the exercises performed in the same Step height showed that both VMO:VLO and VMO:VLL ratios always was greater in the SFS than SFD (p = 0,01), SPS (p = 0,04) and SPD (p = 0,000). There is no difference between SFD and SPS (p = 0,570) or SPD (p = 0,090). In the Control group the SPS was smaller then SPD, on the other hand, in the SDFP group the SPS was greater than SPD (p = 0,30). Our results suggest that there is a difference in the motor control between groups about muscle recruitment either in frontal than posterior Step-up at 75. In the amplitude aspect, the Step at 45 seem to recruit selectively the VMO muscle in comparison with VLL and VLO, thus, this Step height should be used preferentially in the SDFP treatment programs. Regarding of the mode of execution, the frontal Step (SFS) seems to be the most indicated when the objective is the selective activation of the VMO muscle mainly in the Step at 45
Pulzatto Flávio - One of the best experts on this subject based on the ideXlab platform.
-
Atividade elétrica dos músculos estabilizadores da patela em indivíduos portadores da síndrome da dor femoropatelar durante exercícios realizados no Step.
Programa de Pós-Graduação em Fisioterapia - PPGFt, 2005Co-Authors: Pulzatto FlávioAbstract:The purpose of this study was to evaluate the electric activity (EMG) in the temporal and amplitude aspects of the vastus medialis obliquus (VMO), vastus lateralis longus (VLL) and vastus lateralis obliquus (VLO) muscles during forward Step tasks: Stepdown (SFD) and Step-up (SFS), and Backward Step tasks: Step-down (SPD) and Step-up (SPS). Twenty seven females were evaluated and separated in two groups: fifteen normal subjects Control Group (21.13 ± 2.17 years) and twelve subjects with Patellofemoral Pain Syndrome PPS (21.08 ± 2.31 years). The height of the Step was regulated for two angles 45º and 75º - of knee s flexion joint. A metronome was used to help the volunteers about the time to performance the tasks, a eletrogoniometer was used in the control of the knee angle and a pressure sensor was used to determinate the start and the end of the electromyography register. The electric activity was recorded by surface (Ag/AgCl) electrodes, an EMG apparatus with 8 channels (EMG System Brazil) and a software of acquisition data AqDados 7.02. The EMG data was processed by the software Matlab 6.1 that calculated both the onset timing of the muscles and the integrated of the EMG signal. The EMG was normalized by the mean of the three muscle contractions and was calculated de ratio VMO/VLL and VMO/VLO for comparisons between groups and exercises. The relative onset timing was determinate by the difference VMO-VLO and VMO-VLL. The t-tests showed that in the Step up exercise (75º of the knee flexion), there are differences when compared the groups, either in frontal Step VMO/VLL (p= 0.000) and VMO/VLO (p = 0.000), than posterior Step VMO/VLL (p = 0.000) and VMO/VLO (p = 0.000). In the control group there is a prevalence of an early contractions of de VMO muscle, however in the PPS group, the VMO onset occurred at the same time or after the VLO and VLL muscles in the mayor of the cases. The Anova three-way and Duncan post hoc showed that in Step at 45º, the ratio VMO/VLO (p = 0,000) and VMO/VLL (p = 0,016) was greater then Step at 75º in both groups. When compared the Steps in 45º and 75º into the groups, were found differences either in the VMO:VLO ratio (p = 0,000) than VMO:VLL ratio (0,016) with greater values in the Step at 45º. The comparison between the exercises performed in the same Step height showed that both VMO:VLO and VMO:VLL ratios always was greater in the SFS than SFD (p = 0,01), SPS (p = 0,04) and SPD (p = 0,000). There is no difference between SFD and SPS (p = 0,570) or SPD (p = 0,090). In the Control group the SPS was smaller then SPD, on the other hand, in the SDFP group the SPS was greater than SPD (p = 0,30). Our results suggest that there is a difference in the motor control between groups about muscle recruitment either in frontal than posterior Step-up at 75º. In the amplitude aspect, the Step at 45º seem to recruit selectively the VMO muscle in comparison with VLL and VLO, thus, this Step height should be used preferentially in the SDFP treatment programs. Regarding of the mode of execution, the frontal Step (SFS) seems to be the most indicated when the objective is the selective activation of the VMO muscle mainly in the Step at 45º.Universidade Federal de Sao CarlosA proposta deste estudo foi avaliar, nos aspectos temporal e de amplitude, a atividade elétrica dos músculos vasto medial oblíquo (VMO), vasto lateral longo (VLL) e vasto lateral oblíquo (VLO) nos exercícios de Step frontal: subida (SFS) e descida (SFD) e Step posterior: subida (SPS) e descida (SPD). Foram avaliados 27 indivíduos do sexo feminino divididos em dois grupos: 15 clinicamente normais Grupo Controle (21,13 ± 2,17 anos) e 12 portadores da Síndrome da Dor Femoropatelar - SDFP (21,08 ± 2,31). A altura do Step foi regulada para dois ângulos - 45º e 75º - de flexão da articulação do joelho. Um metrônomo auxiliou os voluntários quanto ao tempo de execução do exercício, um eletrogoniômetro foi utilizado para controlar o ângulo de flexão do joelho e um sensor de pressão foi utilizado para informar quanto ao início e o final de cada exercício. A atividade elétrica foi captada por meio de eletrodos ativos diferenciais simples de superfície, um eletromiógrafo de 8 canais (EMG System do Brasil) e um programa de aquisição de dados (AqDados 7.02.06). O sinal elétrico captado foi tratado por rotinas do software Matlab 6.1 que calcularam o tempo de início da ativação elétrica para cada músculo (análise temporal) e a integral matemática da área abaixo da envoltória do sinal retificado e filtrado (análise de amplitude). Os valores da integral foram normalizados pela média das três contrações para cada músculo e posteriormente calculada as relações VMO: VLO e VMO: VLL. O tempo relativo de ativação foi determinado subtraindo-se o tempo de ativação do VLL e do VLO do tempo de ativação do VMO (VMO - VLO e VMO-VLL). O teste t - Student (p ≤ 0,05) revelou que, no exercício de subida no Step a 75º, houve diferença significativa no tempo relativo de ativação entre os grupos, tanto para o Step frontal: VMO-VLO (p = 0,000) e VMO-VLL (p = 0,000), quanto para o Step posterior: VMO-VLO (p = 0,000) e VMO-VLL (p = 0,000). No grupo Controle prevaleceu uma ativação antecipada do VMO em relação aos músculos VLL e VLO; já no grupo SDFP houve prevalência da ativação simultânea e tardia do VMO em relação aos demais músculos. A ANOVA thre-way e o teste de Duncan (p≤ 0,05) revelaram diferenças na relação VMO:VLO e VMO:VLL quando comparados os grupos Controle e SDFP (p = 0,014). Os valores da relação VMO:VLO e VMO:VLL foram significativamente maiores no Step a 45º do que a 75º (p = 0,000 e p = 0,016, respectivamente) nos dois grupos. A comparação entre os exercícios realizados dentro de uma mesma angulação de Step revelou que tanto a relação VMO:VLO quanto VMO:VLL sempre foram maiores no SFS quando comparado ao SFD (p = 0,010), ao SPS (p = 0,040) e ao SPD (p = 0,000). Não houve diferença entre a SFD e a SPS (p=0,570) ou a SPD (p = 0,090). No grupo Controle o SPS foi menor que a SPD (p = 0,030), enquanto que no grupo SDFP ocorreu o inverso nas duas relações em ambos os Steps. Nossos resultados sugerem haver diferenças no controle motor entre os grupos quanto ao recrutamento muscular, tanto no Step frontal quanto no posterior no ângulo de 75º. Na análise da amplitude, o Step a 45 º parece recrutar mais seletivamente o músculo VMO em relação ao VLL e VLO do que no Step a 75º, podendo ser utilizado preferencialmente no tratamento de indivíduos portadores de SDFP. Quanto à modalidade de Step, o exercício de subida frontal (SFS) parece ser o mais indicado quando o objetivo for a ativação seletiva do músculo VMO principalmente no Step a 45º
Nazim Burak Karahanoglu - One of the best experts on this subject based on the ideXlab platform.
-
compressed sensing signal recovery via forward Backward pursuit
Digital Signal Processing, 2013Co-Authors: Nazim Burak Karahanoglu, Hakan ErdoganAbstract:Recovery of sparse signals from compressed measurements constitutes an @?"0 norm minimization problem, which is unpractical to solve. A number of sparse recovery approaches have appeared in the literature, including @?"1 minimization techniques, greedy pursuit algorithms, Bayesian methods and nonconvex optimization techniques among others. This manuscript introduces a novel two stage greedy approach, called the Forward-Backward Pursuit (FBP). FBP is an iterative approach where each iteration consists of consecutive forward and Backward stages. The forward Step first expands the support estimate by the forward Step size, while the following Backward Step shrinks it by the Backward Step size. The forward Step size is larger than the Backward Step size, hence the initially empty support estimate is expanded at the end of each iteration. Forward and Backward Steps are iterated until the residual power of the observation vector falls below a threshold. This structure of FBP does not necessitate the sparsity level to be known a priori in contrast to the Subspace Pursuit or Compressive Sampling Matching Pursuit algorithms. FBP recovery performance is demonstrated via simulations including recovery of random sparse signals with different nonzero coefficient distributions in noisy and noise-free scenarios in addition to the recovery of a sparse image.
-
forward Backward search for compressed sensing signal recovery
European Signal Processing Conference, 2012Co-Authors: Nazim Burak Karahanoglu, Hakan ErdoganAbstract:Reconstruction of sparse signals from reduced dimensions requires the solution of an l 0 norm minimization, which is unpractical. A number of algorithms have appeared in literature, including l 1 minimization, greedy pursuit algorithms, Bayesian methods and nonconvex optimization. This manuscript introduces a greedy approach, called the Forward-Backward Pursuit (FBP), which iteratively enlarges the support by consecutive forward and Backward Steps. At each iteration, the forward Step first expands the support, while the following Backward Step prunes it. The number of atoms selected by the forward Step is selected higher than the number of removals, hence the support is expanded at the end of each iteration. The recovery performance of the proposed method is demonstrated via simulations including different nonzero coefficient distributions in noisy and noise-free scenarios.