Nonlinear transforms have significantly advanced learned image compression(LIC),particularly using residual blocks.This transform enhances the nonlinear expression ability and obtain compact feature representation by ...Nonlinear transforms have significantly advanced learned image compression(LIC),particularly using residual blocks.This transform enhances the nonlinear expression ability and obtain compact feature representation by enlarging the receptive field,which indicates how the convolution process extracts features in a high dimensional feature space.However,its functionality is restricted to the spatial dimension and network depth,limiting further improvements in network performance due to insufficient information interaction and representation.Crucially,the potential of high dimensional feature space in the channel dimension and the exploration of network width/resolution remain largely untapped.In this paper,we consider nonlinear transforms from the perspective of feature space,defining high-dimensional feature spaces in different dimensions and investigating the specific effects.Firstly,we introduce the dimension increasing and decreasing transforms in both channel and spatial dimensions to obtain high dimensional feature space and achieve better feature extraction.Secondly,we design a channel-spatial fusion residual transform(CSR),which incorporates multi-dimensional transforms for a more effective representation.Furthermore,we simplify the proposed fusion transform to obtain a slim architecture(CSR-sm),balancing network complexity and compression performance.Finally,we build the overall network with stacked CSR transforms to achieve better compression and reconstruction.Experimental results demonstrate that the proposed method can achieve superior ratedistortion performance compared to the existing LIC methods and traditional codecs.Specifically,our proposed method achieves 9.38%BD-rate reduction over VVC on Kodak dataset.展开更多
Medical image classification is crucial in disease diagnosis,treatment planning,and clinical decisionmaking.We introduced a novel medical image classification approach that integrates Bayesian Random Semantic Data Aug...Medical image classification is crucial in disease diagnosis,treatment planning,and clinical decisionmaking.We introduced a novel medical image classification approach that integrates Bayesian Random Semantic Data Augmentation(BSDA)with a Vision Mamba-based model for medical image classification(MedMamba),enhanced by residual connection blocks,we named the model BSDA-Mamba.BSDA augments medical image data semantically,enhancing the model’s generalization ability and classification performance.MedMamba,a deep learning-based state space model,excels in capturing long-range dependencies in medical images.By incorporating residual connections,BSDA-Mamba further improves feature extraction capabilities.Through comprehensive experiments on eight medical image datasets,we demonstrate that BSDA-Mamba outperforms existing models in accuracy,area under the curve,and F1-score.Our results highlight BSDA-Mamba’s potential as a reliable tool for medical image analysis,particularly in handling diverse imaging modalities from X-rays to MRI.The open-sourcing of our model’s code and datasets,will facilitate the reproduction and extension of our work.展开更多
Through questionnaires and perceptual maps,we made a variation analysis of landscape spatial image to the people in the Pearl River Delta from the viewpoint of the subjective perception.It is found that the people of ...Through questionnaires and perceptual maps,we made a variation analysis of landscape spatial image to the people in the Pearl River Delta from the viewpoint of the subjective perception.It is found that the people of different ages and different education levels from different areas hold diversified perceptions towards rural landscape space,and they have different ideal rural landscapes.However,there is a common point that the development of rural landscape should retain its excellent local landscape features and the local spirit,which will provide references for building the ideal rural landscape model.展开更多
In this paper,a novel method of ultra-lightweight convolution neural network(CNN)design based on neural architecture search(NAS)and knowledge distillation(KD)is proposed.It can realize the automatic construction of th...In this paper,a novel method of ultra-lightweight convolution neural network(CNN)design based on neural architecture search(NAS)and knowledge distillation(KD)is proposed.It can realize the automatic construction of the space target inverse synthetic aperture radar(ISAR)image recognition model with ultra-lightweight and high accuracy.This method introduces the NAS method into the radar image recognition for the first time,which solves the time-consuming and labor-consuming problems in the artificial design of the space target ISAR image automatic recognition model(STIIARM).On this basis,the NAS model’s knowledge is transferred to the student model with lower computational complexity by the flow of the solution procedure(FSP)distillation method.Thus,the decline of recognition accuracy caused by the direct compression of model structural parameters can be effectively avoided,and the ultralightweight STIIARM can be obtained.In the method,the Inverted Linear Bottleneck(ILB)and Inverted Residual Block(IRB)are firstly taken as each block’s basic structure in CNN.And the expansion ratio,output filter size,number of IRBs,and convolution kernel size are set as the search parameters to construct a hierarchical decomposition search space.Then,the recognition accuracy and computational complexity are taken as the objective function and constraint conditions,respectively,and the global optimization model of the CNN architecture search is established.Next,the simulated annealing(SA)algorithm is used as the search strategy to search out the lightweight and high accuracy STIIARM directly.After that,based on the three principles of similar block structure,the same corresponding channel number,and the minimum computational complexity,the more lightweight student model is designed,and the FSP matrix pairing between the NAS model and student model is completed.Finally,by minimizing the loss between the FSP matrix pairs of the NAS model and student model,the student model’s weight adjustment is completed.Thus the ultra-lightweight and high accuracy STIIARM is obtained.The proposed method’s effectiveness is verified by the simulation experiments on the ISAR image dataset of five types of space targets.展开更多
Mineral dissemination and pore space distribution in ore particles are important features that influence heap leaching performance.To quantify the mineral dissemination and pore space distribution of an ore particle,a...Mineral dissemination and pore space distribution in ore particles are important features that influence heap leaching performance.To quantify the mineral dissemination and pore space distribution of an ore particle,a cylindrical copper oxide ore sample(I center dot 4.6 mm x 5.6 mm)was scanned using high-resolution X-ray computed tomography(HRXCT),a nondestructive imaging technology,at a spatial resolution of 4.85 mu m.Combined with three-dimensional(3D)image analysis techniques,the main mineral phases and pore space were segmented and the volume fraction of each phase was calculated.In addition,the mass fraction of each mineral phase was estimated and the result was validated with that obtained using traditional techniques.Furthermore,the pore phase features,including the pore size distribution,pore surface area,pore fractal dimension,pore centerline,and the pore connectivity,were investigated quantitatively.The pore space analysis results indicate that the pore size distribution closely fits a log-normal distribution and that the pore space morphology is complicated,with a large surface area and low connectivity.This study demonstrates that the combination of HRXCT and 3D image analysis is an effective tool for acquiring 3D mineralogical and pore structural data.展开更多
There exists an increasing need for Milli-Arc-Seconds(MAS)accuracy pointing measurement for current and future space systems.To meet the 0.1″space pointing measurement accuracy requirements of spacecraft in future,th...There exists an increasing need for Milli-Arc-Seconds(MAS)accuracy pointing measurement for current and future space systems.To meet the 0.1″space pointing measurement accuracy requirements of spacecraft in future,the influence of spacecraft micro-vibration on a 0.1″Space Pointing Measuring Instrument(SPMI)is studied.A Quasi-Zero Stiffness Device(QZSD)with adaptive adjustment and variable stroke was proposed.Then,a series of micro-vibration experiments of the SPMI were carried out.The influence of the micro-vibration generated by Guidance Navigation Control(GNC)attitude control components under different attitudes on the SPMI was analyzed.Point spread function of image motion in micro-vibration was also derived.Further,the changes of image motion under the micro-vibration environment were evaluated by extracting the gray centroid of the images,and the experiment processes and results are deeply discussed.The results show that the firstorder frequency of the QZSD system is 0.114 Hz,and it is induced by a double pendulum system;the image motion of single flywheel spinning reached 0.015 pixels;whilst the image motion reached 0.03 pixels when three flywheels are combined spinning.These latest findings provide a beneficial theoretical and technical support for the development of spacecraft with 0·1″pointing accuracy.展开更多
This paper proposes a model for image restoration by combining the wavelet shrinkage and inverse scale space （ISS） method. The ISS is applied to the wavelet representation to modify the retained wavelet coefficients...This paper proposes a model for image restoration by combining the wavelet shrinkage and inverse scale space （ISS） method. The ISS is applied to the wavelet representation to modify the retained wavelet coefficients, and the coefficients smaller than the threshold are set to zero. The curvature term of the ISS can remove the edge artifacts and preserve sharp edges. For the multiscale interpretation of the ISS and the multiscale property of the wavelet representation, small details are preserved. This paper illustrates that the wavelet ISS model can be deduced from the wavelet based on a total variation minimization problem. A stopping criterion is obtained from this minimization in the sense of the Bregman distance in the wavelet domain. Numerical examples show the improvement for the image denoising with the proposed method in the sense of the signal to noise ratio and with fewer details remained in the residue.展开更多
The theoretical framework of visual simulation in virtual reality is discussed. The new concept of visual image space is supposed. On the basis of visual image space, in visual perceptive sense, VR is considered as a ...The theoretical framework of visual simulation in virtual reality is discussed. The new concept of visual image space is supposed. On the basis of visual image space, in visual perceptive sense, VR is considered as a spatial simulation. The objective of the spatial simulation is to transform physical space to visual image space. Last, the prototype system, surveying & mapping virtual Reality (SMVR), is developed, and the space simulation above is realized. By use of SMVR, the real 3D representation, 3D visual analysis, virtual plan and designs can be implemented.展开更多
A support vector regression(SVR) based color image restoration algorithm is proposed.The test color images are firstly mapped into the YUV color space,and then SVR is applied to build up a theoretical model between th...A support vector regression(SVR) based color image restoration algorithm is proposed.The test color images are firstly mapped into the YUV color space,and then SVR is applied to build up a theoretical model between the degraded images and the original one.Performance comparisons of the proposed algorithm versus traditional filtering algorithms are given.Experimental results show that the proposed algorithm has better performance than traditional filtering algorithms and has less computation time than iterative blind deconvolution algorithm.展开更多
Epilepsy is a chronic neurological disorder that affects the function of the brain in people of all ages.It manifests in the electroencephalogram(EEG) signal which records the electrical activity of the brain.Various ...Epilepsy is a chronic neurological disorder that affects the function of the brain in people of all ages.It manifests in the electroencephalogram(EEG) signal which records the electrical activity of the brain.Various image processing,signal processing,and machine-learning based techniques are employed to analyze epilepsy,using spatial and temporal features.The nervous system that generates the EEG signal is considered nonlinear and the EEG signals exhibit chaotic behavior.In order to capture these nonlinear dynamics,we use reconstructed phase space(RPS) representation of the signal.Earlier studies have primarily addressed seizure detection as a binary classification(normal vs.ictal) problem and rarely as a ternary class(normal vs.interictal vs.ictal)problem.We employ transfer learning on a pre-trained deep neural network model and retrain it using RPS images of the EEG signal.The classification accuracy of the model for the binary classes is(98.5±1.5)% and(95±2)% for the ternary classes.The performance of the convolution neural network(CNN) model is better than the other existing statistical approach for all performance indicators such as accuracy,sensitivity,and specificity.The result of the proposed approach shows the prospect of employing RPS images with CNN for predicting epileptic seizures.展开更多
Augmented solar images were used to research the adaptability of four representative image extraction and matching algorithms in space weather domain.These include the scale-invariant feature transform algorithm,speed...Augmented solar images were used to research the adaptability of four representative image extraction and matching algorithms in space weather domain.These include the scale-invariant feature transform algorithm,speeded-up robust features algorithm,binary robust invariant scalable keypoints algorithm,and oriented fast and rotated brief algorithm.The performance of these algorithms was estimated in terms of matching accuracy,feature point richness,and running time.The experiment result showed that no algorithm achieved high accuracy while keeping low running time,and all algorithms are not suitable for image feature extraction and matching of augmented solar images.To solve this problem,an improved method was proposed by using two-frame matching to utilize the accuracy advantage of the scale-invariant feature transform algorithm and the speed advantage of the oriented fast and rotated brief algorithm.Furthermore,our method and the four representative algorithms were applied to augmented solar images.Our application experiments proved that our method achieved a similar high recognition rate to the scale-invariant feature transform algorithm which is significantly higher than other algorithms.Our method also obtained a similar low running time to the oriented fast and rotated brief algorithm,which is significantly lower than other algorithms.展开更多
The Householder transformation-norm structure function in L2 vector space of linear algebra is introduced, and the edge enhancement for remote sensing images is realized. The experiment result is compared with traditi...The Householder transformation-norm structure function in L2 vector space of linear algebra is introduced, and the edge enhancement for remote sensing images is realized. The experiment result is compared with traditional Laplacian and Sobel edge enhancements and it shows that the effect of the new method is better than that of the traditional algorithms.展开更多
In imaging on moving target, it is easy to get space- variant blurred image. In order to recover the image and gain recognizable target, an approach to recover the space-variant blurred image is presented based on ima...In imaging on moving target, it is easy to get space- variant blurred image. In order to recover the image and gain recognizable target, an approach to recover the space-variant blurred image is presented based on image segmentation. Be- cause of motion blur＇s convolution process, the pixels of observed image＇s target and background will be displaced and piled up to produce two superposition regions. As a result, the neighbor- ing pixels in the superposition regions will have similar grey level change. According to the pixel＇s motion-blur character, the target＇s blurred edge of superposition region could be detected. Canny operator can be recurred to detect the target edge which parallels the motion blur direction. Then in the segmentation process, the whole target image which has the character of integral convolution between motion blur and real target image can be obtained. At last, the target image is restored by deconvolution algorithms with adding zeros. The restoration result indicates that the approach can effectively solve the kind of problem of space-variant motion blurred image restoration.展开更多
The parapharyngeal space(PPS) is an inverted pyramid-shaped deep space in the head and neck region, and a variety of tumors, such as salivary gland tumors, neurogenic tumors, nasopharyngeal carcinomas with parapharyng...The parapharyngeal space(PPS) is an inverted pyramid-shaped deep space in the head and neck region, and a variety of tumors, such as salivary gland tumors, neurogenic tumors, nasopharyngeal carcinomas with parapharyngeal invasion, and lymphomas, can be found in this space. The differential diagnosis of PPS tumors remains challenging for radiologists. This study aimed to develop and test a modified method for locating PPS tumors on magnetic resonance(MR) images to improve preoperative differential diagnosis. The new protocol divided the PPS into three compartments: a prestyloid compartment, the carotid sheath, and the areas outside the carotid sheath. PPS tumors were located in these compartments according to the displacements of the tensor veli palatini muscle and the styloid process, with or without blood vessel separations and medial pterygoid invasion. This protocol, as well as a more conventional protocol that is based on displacements of the internal carotid artery(ICA), was used to assess MR images captured from a series of 58 PPS tumors. The consequent distributions of PPS tumor locations determined by both methods were compared. Of all 58 tumors, our new method determined that 57 could be assigned to precise PPS compartments. Nearly all(13/14; 93%) tumors that were located in the pre-styloid compartment were salivary gland tumors. All 15 tumors within the carotid sheath were neurogenic tumors. The vast majority(18/20; 90%) of trans-spatial lesions were malignancies. However, according to the ICA-based method, 28 tumors were located in the pre-styloid compartment, and 24 were located in the post-styloid compartment, leaving 6 tumors that were difficult to locate. Lesions located in both the pre-styloid and the post-styloid compartments comprised various types of tumors. Compared with the conventional ICA-based method, our new method can help radiologists to narrow the differential diagnosis of PPS tumors to specific compartments.展开更多
This paper focuses on the problems of matching a virtual and a real environments by means of hardware and software tools. The real space is represented by a patient’s bone where a set of cuts by means of robot system...This paper focuses on the problems of matching a virtual and a real environments by means of hardware and software tools. The real space is represented by a patient’s bone where a set of cuts by means of robot system is to be made. The virtual space is a 3D model of the bone reconstructed from a set of CT slices. Robot system is then not only to machine bones but also to perform the fundamental step of registration between the two spaces. An external force sensor is used to adjust robot stiffness in order to perform the tactile searching necessary for the registration. A simple but reliable software algorithm is used to control the robot for matching between medical image and robot space in robot-assisted surgery. The results show the system proposed is precise enough for application, and tests been made also clarify the way to improve it.展开更多
It is known that detecting small moving objects in as- tronomical image sequences is a significant research problem in space surveillance. The new theory, compressive sensing, pro- vides a very easy and computationall...It is known that detecting small moving objects in as- tronomical image sequences is a significant research problem in space surveillance. The new theory, compressive sensing, pro- vides a very easy and computationally cheap coding scheme for onboard astronomical remote sensing. An algorithm for small moving space object detection and localization is proposed. The algorithm determines the measurements of objects by comparing the difference between the measurements of the current image and the measurements of the background scene. In contrast to reconstruct the whole image, only a foreground image is recon- structed, which will lead to an effective computational performance, and a high level of localization accuracy is achieved. Experiments and analysis are provided to show the performance of the pro- posed approach on detection and localization.展开更多
In order to implement an observing strategy, image degradation that occurs during optical observation of space debris is ineluctable and has distinct characteris- tics. Image restoration is presented as a way to remov...In order to implement an observing strategy, image degradation that occurs during optical observation of space debris is ineluctable and has distinct characteris- tics. Image restoration is presented as a way to remove the influence of degradation in CCD images of space debris, based on assumed PSF models with the same F-WHM as images of the object. In the process of image restoration, the maximum entropy method is adopted. The results of reduction using observed raw CCD images indi- cate that the precision in estimating positions of objects is improved and the effects of degradation are reduced. Improving the astrometry of space debris using image restoration is effective and feasible.展开更多
This paper advances a three-dimensional space interpolation method of grey / depth image sequence, which breaks free from the limit of original practical photographing route. Pictures can cruise at will in space. By u...This paper advances a three-dimensional space interpolation method of grey / depth image sequence, which breaks free from the limit of original practical photographing route. Pictures can cruise at will in space. By using space sparse sampling, great memorial capacity can be saved and reproduced scenes can be controlled. To solve time consuming and complex computations in three-dimensional interpolation algorithm, we have studied a fast and practical algorithm of scattered space lattice and that of 'Warp' algorithm with proper depth. By several simple aspects of three dimensional space interpolation, we succeed in developing some simple and practical algorithms. Some results of simulated experiments with computers have shown that the new method is absolutely feasible.展开更多
基金supported by the Key Program of the National Natural Science Foundation of China(Grant No.62031013)Guangdong Province Key Construction Discipline Scientific Research Capacity Improvement Project(Grant No.2022ZDJS117).
文摘Nonlinear transforms have significantly advanced learned image compression(LIC),particularly using residual blocks.This transform enhances the nonlinear expression ability and obtain compact feature representation by enlarging the receptive field,which indicates how the convolution process extracts features in a high dimensional feature space.However,its functionality is restricted to the spatial dimension and network depth,limiting further improvements in network performance due to insufficient information interaction and representation.Crucially,the potential of high dimensional feature space in the channel dimension and the exploration of network width/resolution remain largely untapped.In this paper,we consider nonlinear transforms from the perspective of feature space,defining high-dimensional feature spaces in different dimensions and investigating the specific effects.Firstly,we introduce the dimension increasing and decreasing transforms in both channel and spatial dimensions to obtain high dimensional feature space and achieve better feature extraction.Secondly,we design a channel-spatial fusion residual transform(CSR),which incorporates multi-dimensional transforms for a more effective representation.Furthermore,we simplify the proposed fusion transform to obtain a slim architecture(CSR-sm),balancing network complexity and compression performance.Finally,we build the overall network with stacked CSR transforms to achieve better compression and reconstruction.Experimental results demonstrate that the proposed method can achieve superior ratedistortion performance compared to the existing LIC methods and traditional codecs.Specifically,our proposed method achieves 9.38%BD-rate reduction over VVC on Kodak dataset.
文摘Medical image classification is crucial in disease diagnosis,treatment planning,and clinical decisionmaking.We introduced a novel medical image classification approach that integrates Bayesian Random Semantic Data Augmentation(BSDA)with a Vision Mamba-based model for medical image classification(MedMamba),enhanced by residual connection blocks,we named the model BSDA-Mamba.BSDA augments medical image data semantically,enhancing the model’s generalization ability and classification performance.MedMamba,a deep learning-based state space model,excels in capturing long-range dependencies in medical images.By incorporating residual connections,BSDA-Mamba further improves feature extraction capabilities.Through comprehensive experiments on eight medical image datasets,we demonstrate that BSDA-Mamba outperforms existing models in accuracy,area under the curve,and F1-score.Our results highlight BSDA-Mamba’s potential as a reliable tool for medical image analysis,particularly in handling diverse imaging modalities from X-rays to MRI.The open-sourcing of our model’s code and datasets,will facilitate the reproduction and extension of our work.
文摘Through questionnaires and perceptual maps,we made a variation analysis of landscape spatial image to the people in the Pearl River Delta from the viewpoint of the subjective perception.It is found that the people of different ages and different education levels from different areas hold diversified perceptions towards rural landscape space,and they have different ideal rural landscapes.However,there is a common point that the development of rural landscape should retain its excellent local landscape features and the local spirit,which will provide references for building the ideal rural landscape model.
文摘In this paper,a novel method of ultra-lightweight convolution neural network(CNN)design based on neural architecture search(NAS)and knowledge distillation(KD)is proposed.It can realize the automatic construction of the space target inverse synthetic aperture radar(ISAR)image recognition model with ultra-lightweight and high accuracy.This method introduces the NAS method into the radar image recognition for the first time,which solves the time-consuming and labor-consuming problems in the artificial design of the space target ISAR image automatic recognition model(STIIARM).On this basis,the NAS model’s knowledge is transferred to the student model with lower computational complexity by the flow of the solution procedure(FSP)distillation method.Thus,the decline of recognition accuracy caused by the direct compression of model structural parameters can be effectively avoided,and the ultralightweight STIIARM can be obtained.In the method,the Inverted Linear Bottleneck(ILB)and Inverted Residual Block(IRB)are firstly taken as each block’s basic structure in CNN.And the expansion ratio,output filter size,number of IRBs,and convolution kernel size are set as the search parameters to construct a hierarchical decomposition search space.Then,the recognition accuracy and computational complexity are taken as the objective function and constraint conditions,respectively,and the global optimization model of the CNN architecture search is established.Next,the simulated annealing(SA)algorithm is used as the search strategy to search out the lightweight and high accuracy STIIARM directly.After that,based on the three principles of similar block structure,the same corresponding channel number,and the minimum computational complexity,the more lightweight student model is designed,and the FSP matrix pairing between the NAS model and student model is completed.Finally,by minimizing the loss between the FSP matrix pairs of the NAS model and student model,the student model’s weight adjustment is completed.Thus the ultra-lightweight and high accuracy STIIARM is obtained.The proposed method’s effectiveness is verified by the simulation experiments on the ISAR image dataset of five types of space targets.
基金financially supported by the National Natural Science Foundation of China(No.51304076)the Natural Science Foundation of Hunan Province,China(No.14JJ4064)
文摘Mineral dissemination and pore space distribution in ore particles are important features that influence heap leaching performance.To quantify the mineral dissemination and pore space distribution of an ore particle,a cylindrical copper oxide ore sample(I center dot 4.6 mm x 5.6 mm)was scanned using high-resolution X-ray computed tomography(HRXCT),a nondestructive imaging technology,at a spatial resolution of 4.85 mu m.Combined with three-dimensional(3D)image analysis techniques,the main mineral phases and pore space were segmented and the volume fraction of each phase was calculated.In addition,the mass fraction of each mineral phase was estimated and the result was validated with that obtained using traditional techniques.Furthermore,the pore phase features,including the pore size distribution,pore surface area,pore fractal dimension,pore centerline,and the pore connectivity,were investigated quantitatively.The pore space analysis results indicate that the pore size distribution closely fits a log-normal distribution and that the pore space morphology is complicated,with a large surface area and low connectivity.This study demonstrates that the combination of HRXCT and 3D image analysis is an effective tool for acquiring 3D mineralogical and pore structural data.
基金the support from the National Natural Science Foundation of China(No.51905034,52275083)。
文摘There exists an increasing need for Milli-Arc-Seconds(MAS)accuracy pointing measurement for current and future space systems.To meet the 0.1″space pointing measurement accuracy requirements of spacecraft in future,the influence of spacecraft micro-vibration on a 0.1″Space Pointing Measuring Instrument(SPMI)is studied.A Quasi-Zero Stiffness Device(QZSD)with adaptive adjustment and variable stroke was proposed.Then,a series of micro-vibration experiments of the SPMI were carried out.The influence of the micro-vibration generated by Guidance Navigation Control(GNC)attitude control components under different attitudes on the SPMI was analyzed.Point spread function of image motion in micro-vibration was also derived.Further,the changes of image motion under the micro-vibration environment were evaluated by extracting the gray centroid of the images,and the experiment processes and results are deeply discussed.The results show that the firstorder frequency of the QZSD system is 0.114 Hz,and it is induced by a double pendulum system;the image motion of single flywheel spinning reached 0.015 pixels;whilst the image motion reached 0.03 pixels when three flywheels are combined spinning.These latest findings provide a beneficial theoretical and technical support for the development of spacecraft with 0·1″pointing accuracy.
基金supported by the National Natural Science Foundation of China (61101208)
文摘This paper proposes a model for image restoration by combining the wavelet shrinkage and inverse scale space （ISS） method. The ISS is applied to the wavelet representation to modify the retained wavelet coefficients, and the coefficients smaller than the threshold are set to zero. The curvature term of the ISS can remove the edge artifacts and preserve sharp edges. For the multiscale interpretation of the ISS and the multiscale property of the wavelet representation, small details are preserved. This paper illustrates that the wavelet ISS model can be deduced from the wavelet based on a total variation minimization problem. A stopping criterion is obtained from this minimization in the sense of the Bregman distance in the wavelet domain. Numerical examples show the improvement for the image denoising with the proposed method in the sense of the signal to noise ratio and with fewer details remained in the residue.
文摘The theoretical framework of visual simulation in virtual reality is discussed. The new concept of visual image space is supposed. On the basis of visual image space, in visual perceptive sense, VR is considered as a spatial simulation. The objective of the spatial simulation is to transform physical space to visual image space. Last, the prototype system, surveying & mapping virtual Reality (SMVR), is developed, and the space simulation above is realized. By use of SMVR, the real 3D representation, 3D visual analysis, virtual plan and designs can be implemented.
基金the National Natural Science Foundation of China (No. 60675023)
文摘A support vector regression(SVR) based color image restoration algorithm is proposed.The test color images are firstly mapped into the YUV color space,and then SVR is applied to build up a theoretical model between the degraded images and the original one.Performance comparisons of the proposed algorithm versus traditional filtering algorithms are given.Experimental results show that the proposed algorithm has better performance than traditional filtering algorithms and has less computation time than iterative blind deconvolution algorithm.
文摘Epilepsy is a chronic neurological disorder that affects the function of the brain in people of all ages.It manifests in the electroencephalogram(EEG) signal which records the electrical activity of the brain.Various image processing,signal processing,and machine-learning based techniques are employed to analyze epilepsy,using spatial and temporal features.The nervous system that generates the EEG signal is considered nonlinear and the EEG signals exhibit chaotic behavior.In order to capture these nonlinear dynamics,we use reconstructed phase space(RPS) representation of the signal.Earlier studies have primarily addressed seizure detection as a binary classification(normal vs.ictal) problem and rarely as a ternary class(normal vs.interictal vs.ictal)problem.We employ transfer learning on a pre-trained deep neural network model and retrain it using RPS images of the EEG signal.The classification accuracy of the model for the binary classes is(98.5±1.5)% and(95±2)% for the ternary classes.The performance of the convolution neural network(CNN) model is better than the other existing statistical approach for all performance indicators such as accuracy,sensitivity,and specificity.The result of the proposed approach shows the prospect of employing RPS images with CNN for predicting epileptic seizures.
基金Supported by the Key Research Program of the Chinese Academy of Sciences(ZDRE-KT-2021-3)。
文摘Augmented solar images were used to research the adaptability of four representative image extraction and matching algorithms in space weather domain.These include the scale-invariant feature transform algorithm,speeded-up robust features algorithm,binary robust invariant scalable keypoints algorithm,and oriented fast and rotated brief algorithm.The performance of these algorithms was estimated in terms of matching accuracy,feature point richness,and running time.The experiment result showed that no algorithm achieved high accuracy while keeping low running time,and all algorithms are not suitable for image feature extraction and matching of augmented solar images.To solve this problem,an improved method was proposed by using two-frame matching to utilize the accuracy advantage of the scale-invariant feature transform algorithm and the speed advantage of the oriented fast and rotated brief algorithm.Furthermore,our method and the four representative algorithms were applied to augmented solar images.Our application experiments proved that our method achieved a similar high recognition rate to the scale-invariant feature transform algorithm which is significantly higher than other algorithms.Our method also obtained a similar low running time to the oriented fast and rotated brief algorithm,which is significantly lower than other algorithms.
基金Funded by the National Natural Science Foundation of China(No.40571100).
文摘The Householder transformation-norm structure function in L2 vector space of linear algebra is introduced, and the edge enhancement for remote sensing images is realized. The experiment result is compared with traditional Laplacian and Sobel edge enhancements and it shows that the effect of the new method is better than that of the traditional algorithms.
文摘In imaging on moving target, it is easy to get space- variant blurred image. In order to recover the image and gain recognizable target, an approach to recover the space-variant blurred image is presented based on image segmentation. Be- cause of motion blur＇s convolution process, the pixels of observed image＇s target and background will be displaced and piled up to produce two superposition regions. As a result, the neighbor- ing pixels in the superposition regions will have similar grey level change. According to the pixel＇s motion-blur character, the target＇s blurred edge of superposition region could be detected. Canny operator can be recurred to detect the target edge which parallels the motion blur direction. Then in the segmentation process, the whole target image which has the character of integral convolution between motion blur and real target image can be obtained. At last, the target image is restored by deconvolution algorithms with adding zeros. The restoration result indicates that the approach can effectively solve the kind of problem of space-variant motion blurred image restoration.
文摘The parapharyngeal space(PPS) is an inverted pyramid-shaped deep space in the head and neck region, and a variety of tumors, such as salivary gland tumors, neurogenic tumors, nasopharyngeal carcinomas with parapharyngeal invasion, and lymphomas, can be found in this space. The differential diagnosis of PPS tumors remains challenging for radiologists. This study aimed to develop and test a modified method for locating PPS tumors on magnetic resonance(MR) images to improve preoperative differential diagnosis. The new protocol divided the PPS into three compartments: a prestyloid compartment, the carotid sheath, and the areas outside the carotid sheath. PPS tumors were located in these compartments according to the displacements of the tensor veli palatini muscle and the styloid process, with or without blood vessel separations and medial pterygoid invasion. This protocol, as well as a more conventional protocol that is based on displacements of the internal carotid artery(ICA), was used to assess MR images captured from a series of 58 PPS tumors. The consequent distributions of PPS tumor locations determined by both methods were compared. Of all 58 tumors, our new method determined that 57 could be assigned to precise PPS compartments. Nearly all(13/14; 93%) tumors that were located in the pre-styloid compartment were salivary gland tumors. All 15 tumors within the carotid sheath were neurogenic tumors. The vast majority(18/20; 90%) of trans-spatial lesions were malignancies. However, according to the ICA-based method, 28 tumors were located in the pre-styloid compartment, and 24 were located in the post-styloid compartment, leaving 6 tumors that were difficult to locate. Lesions located in both the pre-styloid and the post-styloid compartments comprised various types of tumors. Compared with the conventional ICA-based method, our new method can help radiologists to narrow the differential diagnosis of PPS tumors to specific compartments.
文摘This paper focuses on the problems of matching a virtual and a real environments by means of hardware and software tools. The real space is represented by a patient’s bone where a set of cuts by means of robot system is to be made. The virtual space is a 3D model of the bone reconstructed from a set of CT slices. Robot system is then not only to machine bones but also to perform the fundamental step of registration between the two spaces. An external force sensor is used to adjust robot stiffness in order to perform the tactile searching necessary for the registration. A simple but reliable software algorithm is used to control the robot for matching between medical image and robot space in robot-assisted surgery. The results show the system proposed is precise enough for application, and tests been made also clarify the way to improve it.
基金supported by the National Natural Science Foundation of China (60903126)the China Postdoctoral Special Science Foundation (201003685)+1 种基金the China Postdoctoral Science Foundation (20090451397)the Northwestern Polytechnical University Foundation for Fundamental Research (JC201120)
文摘It is known that detecting small moving objects in as- tronomical image sequences is a significant research problem in space surveillance. The new theory, compressive sensing, pro- vides a very easy and computationally cheap coding scheme for onboard astronomical remote sensing. An algorithm for small moving space object detection and localization is proposed. The algorithm determines the measurements of objects by comparing the difference between the measurements of the current image and the measurements of the background scene. In contrast to reconstruct the whole image, only a foreground image is recon- structed, which will lead to an effective computational performance, and a high level of localization accuracy is achieved. Experiments and analysis are provided to show the performance of the pro- posed approach on detection and localization.
基金funded by the National Natural Science Foundation of China (Grant Nos.11125315 and 11033009)
文摘In order to implement an observing strategy, image degradation that occurs during optical observation of space debris is ineluctable and has distinct characteris- tics. Image restoration is presented as a way to remove the influence of degradation in CCD images of space debris, based on assumed PSF models with the same F-WHM as images of the object. In the process of image restoration, the maximum entropy method is adopted. The results of reduction using observed raw CCD images indi- cate that the precision in estimating positions of objects is improved and the effects of degradation are reduced. Improving the astrometry of space debris using image restoration is effective and feasible.
文摘This paper advances a three-dimensional space interpolation method of grey / depth image sequence, which breaks free from the limit of original practical photographing route. Pictures can cruise at will in space. By using space sparse sampling, great memorial capacity can be saved and reproduced scenes can be controlled. To solve time consuming and complex computations in three-dimensional interpolation algorithm, we have studied a fast and practical algorithm of scattered space lattice and that of 'Warp' algorithm with proper depth. By several simple aspects of three dimensional space interpolation, we succeed in developing some simple and practical algorithms. Some results of simulated experiments with computers have shown that the new method is absolutely feasible.
