The popularly used circulant matrix model of deconvolution is mostly heavily ill-posed or singular and it is not suitable to many blind deconvolution problems. The aperiodic matrix model can improve the condition numb...The popularly used circulant matrix model of deconvolution is mostly heavily ill-posed or singular and it is not suitable to many blind deconvolution problems. The aperiodic matrix model can improve the condition number of deconvolution problems and its accommodation is much wider than the circulant one's. This paper discusses a comparison of the two models including their ill-posedness, the rationality of the approximation by the models, and their computational efficiency. The comparison shows that the aperiodic model is promising in the development of new restoration algorithms.展开更多
For the multisensor linear discrete time-invariant stochastic systems with correlated noises and unknown noise statistics,an on-line noise statistics estimator is presented by using the correlation method.Substituting...For the multisensor linear discrete time-invariant stochastic systems with correlated noises and unknown noise statistics,an on-line noise statistics estimator is presented by using the correlation method.Substituting it into the steady-state Riccati equation,the self-tuning Riccati equation is obtained.Using the Kalman filtering method,based on the self-tuning Riccati equation,a self-tuning weighted measurement fusion white noise deconvolution estimator is presented.By the dynamic error system analysis（DESA） method,it is proved that the self-tuning fusion white noise deconvolution estimator converges to the optimal fusion steadystate white noise deconvolution estimator in a realization,so that it has the asymptotic global optimality.A simulation example for Bernoulli-Gaussian input white noise shows its effectiveness.展开更多
A deconvolution algorithm is proposed to account for the distortions of impulse shape introduced by propagation process. By finding the best correlation of the received waveform with the multiple templates, the number...A deconvolution algorithm is proposed to account for the distortions of impulse shape introduced by propagation process. By finding the best correlation of the received waveform with the multiple templates, the number of multipath components is reduced as the result of eliminating the ＂phantom paths＂, and the captured energy increases. Moreover, it needs only a single reference measurement in real measurement environment （do not need the anechoic chamber）, which by far simplifies the templates acquiring procedure.展开更多
Here we present a simple yet effective gas chromatography-mass spectrometry(GC-MS)identification approach for the detection of heteroatom-containing compounds(HACCs)in petroleum fractions.The MS/AMDIS(Automated Mass S...Here we present a simple yet effective gas chromatography-mass spectrometry(GC-MS)identification approach for the detection of heteroatom-containing compounds(HACCs)in petroleum fractions.The MS/AMDIS(Automated Mass Spectral Deconvolution and Identification System)program was used to identify parts per million(ppm)HACC concentrations in petroleum fractions in place of traditional techniques(extraction and standard injection).Polycyclic aromatic sulfur heterocycles(S-PAHs)were used as model compounds to confirm the validity of the AMDIS identifiers,which were compared with extracted results using the off-line X-calibur software.AMDIS was able to identify ppm concentrations of S-PAHs in oil condensate.There was good agreement between experimental and AMDIS identification results for S-PAHs in oil condensate.AMDIS was also used to detect nitrogen-containing compounds(NCCs)and alkylphenols in oil condensate.Our results confirmed the presence of 2-methylbenzothiazole,carbazole,and 2,4-ditertbutyl phenol.In a crude oil sample,AMDIS identification of m/z=191 biomarkers wa s consistent with empirical results.Therefore,AMDIS can help to reduce the number of experimental steps in identification protocols.展开更多
An algorithm is presented for image prior combinations based blind deconvolution and applied to astronomical images.Using a hierarchical Bayesian framework, the unknown original image and all required algorithmic para...An algorithm is presented for image prior combinations based blind deconvolution and applied to astronomical images.Using a hierarchical Bayesian framework, the unknown original image and all required algorithmic parameters are estimated simultaneously. Through utilization of variational Bayesian analysis,approximations of the posterior distributions on each unknown are obtained by minimizing the Kullback-Leibler(KL) distance, thus providing uncertainties of the estimates during the restoration process. Experimental results on both synthetic images and real astronomical images demonstrate that the proposed approaches compare favorably to other state-of-the-art reconstruction methods.展开更多
While moving ahead with the object detection technology, especially deep neural networks, many related tasks, such as medical application and industrial automation, have achieved great success. However, the detection ...While moving ahead with the object detection technology, especially deep neural networks, many related tasks, such as medical application and industrial automation, have achieved great success. However, the detection of objects with multiple aspect ratios and scales is still a key problem. This paper proposes a top-down and bottom-up feature pyramid network(TDBU-FPN),which combines multi-scale feature representation and anchor generation at multiple aspect ratios. First, in order to build the multi-scale feature map, this paper puts a number of fully convolutional layers after the backbone. Second, to link neighboring feature maps, top-down and bottom-up flows are adopted to introduce context information via top-down flow and supplement suboriginal information via bottom-up flow. The top-down flow refers to the deconvolution procedure, and the bottom-up flow refers to the pooling procedure. Third, the problem of adapting different object aspect ratios is tackled via many anchor shapes with different aspect ratios on each multi-scale feature map. The proposed method is evaluated on the pattern analysis, statistical modeling and computational learning visual object classes(PASCAL VOC)dataset and reaches an accuracy of 79%, which exhibits a 1.8% improvement with a detection speed of 23 fps.展开更多
文摘The popularly used circulant matrix model of deconvolution is mostly heavily ill-posed or singular and it is not suitable to many blind deconvolution problems. The aperiodic matrix model can improve the condition number of deconvolution problems and its accommodation is much wider than the circulant one's. This paper discusses a comparison of the two models including their ill-posedness, the rationality of the approximation by the models, and their computational efficiency. The comparison shows that the aperiodic model is promising in the development of new restoration algorithms.
基金supported by the National Natural Science Foundation of China(60874063)Science and Technology Research Foundation of Heilongjiang Education Department(11551355)Key Laboratory of Electronics Engineering,College of Heilongjiang Province(DZZD20105)
文摘For the multisensor linear discrete time-invariant stochastic systems with correlated noises and unknown noise statistics,an on-line noise statistics estimator is presented by using the correlation method.Substituting it into the steady-state Riccati equation,the self-tuning Riccati equation is obtained.Using the Kalman filtering method,based on the self-tuning Riccati equation,a self-tuning weighted measurement fusion white noise deconvolution estimator is presented.By the dynamic error system analysis（DESA） method,it is proved that the self-tuning fusion white noise deconvolution estimator converges to the optimal fusion steadystate white noise deconvolution estimator in a realization,so that it has the asymptotic global optimality.A simulation example for Bernoulli-Gaussian input white noise shows its effectiveness.
基金the Key Program of the National Natural Science Foundation of China (60432040)the China Postdoctors Science Foundation (20060390792).
文摘A deconvolution algorithm is proposed to account for the distortions of impulse shape introduced by propagation process. By finding the best correlation of the received waveform with the multiple templates, the number of multipath components is reduced as the result of eliminating the ＂phantom paths＂, and the captured energy increases. Moreover, it needs only a single reference measurement in real measurement environment （do not need the anechoic chamber）, which by far simplifies the templates acquiring procedure.
文摘Here we present a simple yet effective gas chromatography-mass spectrometry(GC-MS)identification approach for the detection of heteroatom-containing compounds(HACCs)in petroleum fractions.The MS/AMDIS(Automated Mass Spectral Deconvolution and Identification System)program was used to identify parts per million(ppm)HACC concentrations in petroleum fractions in place of traditional techniques(extraction and standard injection).Polycyclic aromatic sulfur heterocycles(S-PAHs)were used as model compounds to confirm the validity of the AMDIS identifiers,which were compared with extracted results using the off-line X-calibur software.AMDIS was able to identify ppm concentrations of S-PAHs in oil condensate.There was good agreement between experimental and AMDIS identification results for S-PAHs in oil condensate.AMDIS was also used to detect nitrogen-containing compounds(NCCs)and alkylphenols in oil condensate.Our results confirmed the presence of 2-methylbenzothiazole,carbazole,and 2,4-ditertbutyl phenol.In a crude oil sample,AMDIS identification of m/z=191 biomarkers wa s consistent with empirical results.Therefore,AMDIS can help to reduce the number of experimental steps in identification protocols.
文摘An algorithm is presented for image prior combinations based blind deconvolution and applied to astronomical images.Using a hierarchical Bayesian framework, the unknown original image and all required algorithmic parameters are estimated simultaneously. Through utilization of variational Bayesian analysis,approximations of the posterior distributions on each unknown are obtained by minimizing the Kullback-Leibler(KL) distance, thus providing uncertainties of the estimates during the restoration process. Experimental results on both synthetic images and real astronomical images demonstrate that the proposed approaches compare favorably to other state-of-the-art reconstruction methods.
基金supported by the Program of Introducing Talents of Discipline to Universities(111 Plan)of China(B14010)the National Natural Science Foundation of China(31727901)
文摘While moving ahead with the object detection technology, especially deep neural networks, many related tasks, such as medical application and industrial automation, have achieved great success. However, the detection of objects with multiple aspect ratios and scales is still a key problem. This paper proposes a top-down and bottom-up feature pyramid network(TDBU-FPN),which combines multi-scale feature representation and anchor generation at multiple aspect ratios. First, in order to build the multi-scale feature map, this paper puts a number of fully convolutional layers after the backbone. Second, to link neighboring feature maps, top-down and bottom-up flows are adopted to introduce context information via top-down flow and supplement suboriginal information via bottom-up flow. The top-down flow refers to the deconvolution procedure, and the bottom-up flow refers to the pooling procedure. Third, the problem of adapting different object aspect ratios is tackled via many anchor shapes with different aspect ratios on each multi-scale feature map. The proposed method is evaluated on the pattern analysis, statistical modeling and computational learning visual object classes(PASCAL VOC)dataset and reaches an accuracy of 79%, which exhibits a 1.8% improvement with a detection speed of 23 fps.
