- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
PET Image Reconstruction Using Deep Image Prior
摘要： Recently deep neural networks have been widely and successfully applied in computer vision tasks and attracted growing interests in medical imaging. One barrier for the application of deep neural networks to medical imaging is the need of large amounts of prior training pairs, which is not always feasible in clinical practice. This is especially true for medical image reconstruction problems, where raw data are needed. Inspired by the deep image prior framework, in this work we proposed a personalized network training method where no prior training pairs are needed, but only the patient’ own prior information. The network is updated during the iterative reconstruction process using the patient specific prior information and measured data. We formulated the maximum likelihood estimation as a constrained optimization problem and solved it using the alternating direction method of multipliers (ADMM) algorithm. Magnetic resonance imaging (MRI) guided Positron emission tomography (PET) reconstruction was employed as an example to demonstrate the effectiveness of the proposed framework. Quantification results based on simulation and real data show that the proposed reconstruction framework can outperform Gaussian post-smoothing and anatomically-guided reconstructions using the kernel method or the neural network penalty.
关键词： positron emission tomography,unsupervised learning,Medical image reconstruction,deep neural network
更新于2025-09-04 15:30:14
-
[IEEE 2018 25th IEEE International Conference on Image Processing (ICIP) - Athens, Greece (2018.10.7-2018.10.10)] 2018 25th IEEE International Conference on Image Processing (ICIP) - Improving Time-of-Flight Sensor for Specular Surfaces with Shape from Polarization
摘要： Time-of-Flight (ToF) sensors can obtain depth values for diffuse objects. However, the essential problem is that these sensors cannot receive active light from specular surfaces due to specular reflections. In this paper, we propose a new depth reconstruction framework for specular objects that combines ToF cues and Shape from Polarization (SfP). To overcome the ill-posedness of SfP with a single view, we integrate superpixel segmentation with planarity constraints for every superpixel. Experimental results demonstrate the effectiveness of the depth reconstruction algorithm for both controlled environment data and real vehicle data in a parking area.
关键词： Shape from Polarization,specular surface,Time-of-Flight sensor,single view reconstruction,superpixel segmentation
更新于2025-09-04 15:30:14
-
Dose reconstruction from PET images in carbon ion therapy: a deconvolution approach
摘要： Dose and range verification have become important tools to bring carbon ion therapy to a higher level of confidence in clinical applications. Positron emission tomography is among the most commonly used approaches for this purpose and relies on the creation of positron emitting nuclei in nuclear interactions of the primary ions with tissue. Predictions of these positron emitter distributions are usually obtained from time-consuming Monte Carlo simulations or measurements from previous treatment fractions, and their comparison to the current, measured image allows for treatment verification. Still, a direct comparison of planned and delivered dose would be highly desirable, since the dose is the quantity of interest in radiation therapy and its confirmation improves quality assurance in carbon ion therapy. In this work, we present a deconvolution approach to predict dose distributions from PET images in carbon ion therapy. Under the assumption that the one-dimensional PET distribution is described by a convolution of the depth dose distribution and a filter kernel, an evolutionary algorithm is introduced to perform the reverse step and predict the depth dose distribution from a measured PET distribution. Filter kernels are obtained from either a library or are created for any given situation on-the-fly, using predictions of the β+-decay and depth dose distributions, and the very same evolutionary algorithm. The applicability of this approach is demonstrated for monoenergetic and polyenergetic carbon ion irradiation of homogeneous and heterogeneous solid phantoms as well as a patient computed tomography image, using Monte Carlo simulated distributions and measured in-beam PET data. Carbon ion ranges are predicted within less than 0.5 mm and 1 mm deviation for simulated and measured distributions, respectively.
关键词： evolutionary algorithm,PET imaging,range verification,carbon ion therapy,dose reconstruction
更新于2025-09-04 15:30:14
-
Study of Sensitivity and Resolution for Full Ring PET Prototypes based on Continuous Crystals and analytical modeling of the light distribution
摘要： Sensitivity and spatial resolution are the main parameters to maximize in the performance of a PET scanner. For this purpose, detectors consisting of a combination of continuous crystals optically coupled to segmented photodetectors have been employed. With the use of continuous crystals the sensitivity is increased with respect to the pixelated crystals. In addition, spatial resolution is no longer limited to the crystal size. The main drawback is the difficulty in determining the interaction position. In this work, we present the characterization of the performance of a full ring based on cuboid continuous crystals coupled to SiPMs. To this end, we have employed the simulations developed in a previous work for our experimental detector head. Sensitivity could be further enhanced by using tapered crystals. This enhancement is obtained by increasing the solid angle coverage, reducing the wedge-shaped gaps between contiguous detectors. The performance of the scanners based on both crystal geometries was characterized following NEMA NU 4-2008 standardized protocol in order to compare them. An average sensitivity gain over the entire axial field of view of 13.63% has been obtained with tapered geometry while similar performance of the spatial resolution has been proven with both scanners. The activity at which NECR and True peak occur is smaller and the peak value is greater for tapered crystals than for cuboid crystals. Moreover, a higher degree of homogeneity was obtained in the sensitivity map due to the tighter packing of the crystals, which reduces the gaps and results in a better recovery of homogeneous regions than for the cuboid configuration. Some of the results obtained, such as spatial resolution, depend on the interaction position estimation and may vary if other method is employed.
关键词： NEMA NU 4-2008,Monte Carlo simulations,image reconstruction,continuous crystals,depth of interaction,positron emission tomography (PET)
更新于2025-09-04 15:30:14
-
Approximate quantum state reconstruction without a quantum channel
摘要： We investigate the optimal quantum state reconstruction from the cloud to many spatially separated users by a measure-broadcast-prepare scheme without the availability of the quantum channel. The quantum state equally distributed from the cloud to an arbitrary number of users is generated at each port by an ensemble of known quantum states with assistance from classical information of measurement outcomes by broadcasting. The obtained quantum state for each user is optimal in the sense that the ?delity universally achieves the upper bound. We present the universal quantum state distribution by providing physical realizable measurement bases in the cloud as well as the reconstruction method for each user. The quantum state reconstruction scheme works for arbitrary many identical pure input states in the general dimensional system.
关键词： measure-broadcast-prepare scheme,quantum channel,quantum state reconstruction,optimal ?delity
更新于2025-09-04 15:30:14
-
Iterative Reweighted Tikhonov-regularized Multihypothesis Prediction Scheme for Distributed Compressive Video Sensing
摘要： Distributed compressive video sensing (DCVS) has great potential for signal acquisition and processing in source-limited communication, e.g., wireless video sensors network (WVSN), because it shifts complicated motion estimation and motion compensation from the encoder to the decoder. Known as a state-of-the-art technique in DCVS, multihypothesis (MH) prediction is widely used because of its acceptable performance and low computational complexity. However, this technique is restricted by inaccurate regularizations, which can cause susceptibility to inaccurate hypotheses. In this paper, we present an iterative reweighted Tikhonov-regularized scheme for MH prediction reconstruction. Specifically, to enhance robustness, this scheme proposes a reweighted Tikhonov regularization (MH-RTIK) that synthetically considers three factors that affect MH prediction performance—accuracy of the hypothesis set, number of hypotheses, and accuracy of regularizations—by utilizing the influence of each hypothesis. Furthermore, to avoid over-iteration in iterative MH prediction reconstruction, we propose a Bhattacharyya coefficient-based stopping criterion for use in the recovery of non-key frames, in which we exploit the similarity to an adjacent key frame rather than a previous iteration result. The simulation results show that the proposed scheme outperforms the state-of-the-art MH methods in terms of robustness to inaccurate hypotheses when there are a limited number of hypotheses.
关键词： Distributed compressive video sensing (DCVS),video reconstruction,wireless video sensors network (WVSN),multihypothesis (MH) prediction
更新于2025-09-04 15:30:14
-
Image reconstruction for frequency-domain diffuse optical tomography
摘要： The image reconstruction algorithm of diffuse optical tomography (DOT) is based on the diffusion equation and involves both the forward problem and inverse solution. The forward problem solves the diffusion equation using the finite element method for calculating the transmitted light distribution under the condition of presumed light source and optical coefficient. The inverse solution reconstructs the optical property coefficient distribution using Newton’s method. The work within this study develops an image reconstruction algorithm for frequency-domain DOT. A numerical simulations approach to light propagation in the tissue is conducted, while the optical property is reconstructed employing data around the boundary. We implement different designated simulation cases, including different contrast ratios of absorption and reduced scattering coefficient of inclusion with respect to the background used for verifying the results of the forward problem and the developed reconstruction algorithm. Reconstruction results indicate that the quality of reconstructed images can be effective for screening breast cancer.
关键词： frequency domain,Diffuse optical tomography,image reconstruction
更新于2025-09-04 15:30:14
-
Exploring RGB-D Cameras for 3D Reconstruction of Cultural Heritage
摘要： RGB-D cameras have a great potential to solve several problems arising during the digitization of objects, such as cultural heritage. Three-dimensional (3D) digital preservation is usually performed with the use of high-end 3D scanners, as the 3D points generated by this type of equipment are in average millimeter up to sub-millimeter accurate. The downside of 3D scanners, in addition to the high cost, is the infrastructure requirements. It requires its own source of energy, a large workspace with tripods, special training to calibrate and operate the equipment, and high acquisition time, potentially taking several minutes for capturing a single image. An alternative is the use of low-cost depth cameras that are easy to operate and only require connection to a laptop and a source of energy. There are several recent studies showing the potential of RGB-D sensors. However, they often exhibit errors when applied to a full 360 degrees 3D reconstruction setup, known as the loop closure problem. This kind of error accumulation is intensified by the lower accuracy and large volume of data generated by RGB-D cameras. This article proposes a complete methodology for 3D reconstruction based on RGB-D sensors. To mitigate the loop closure effect, a pairwise alignment method was developed. The proposed approach expands the connectivity graph connections in a pairwise alignment system, by automatically discovering new pairs of meshes with overlapping regions. Then the alignment is more evenly distributed over the aligned pairs, avoiding the loop closure problem of full 3D reconstructions. The experiments were performed on a collection of 30 artworks made by the Baroque artist Antonio Francisco Lisboa, known as Aleijadinho, as part of the Aleijadinho Digital project conducted in partnership with IPHAN (Brazilian National Institute for Cultural and Artistic Heritage) and United Nations Educational, Scientific and Cultural Organization (UNESCO). Experimental results show 3D models that are favorably compared to state-of-the-art methods available in the literature using RGD-D sensors. The main contributions of this work are: a new method for 3D alignment dedicated to attenuate the RGB-D camera loop closure problem; the development and disclosure of a complete, practical solution for 3D reconstruction of artworks; and the construction of 3D digital models of an important and challenging collection of Brazilian cultural heritage, made accessible by a virtual museum.
关键词： global registration,cultural heritage,3D reconstruction
更新于2025-09-04 15:30:14
-
Example-based image super-resolution via blur kernel estimation and variational reconstruction
摘要： Single image super-resolution aims at generating clear high-resolution image from one low-resolution image. Due to the limited low-resolution information, it is a challenging task to restore clear, artifacts-free image, meanwhile preserving finer structures and textures. This paper proposes an effective example-based image super-resolution method while making clear image and no compromise on quality. Firstly, the image prior is imposed on the anchor neighborhood regression model to optimize mapping coefficient for interim latent image construction. In order to remove its blur, kernel estimation iteration optimization algorithm is proposed based on the salient edges which are extracted through texture-structure discriminate minimum energy function and fractional order mask enhancement. Finally, an accurate reconstruction constraint combined with a simple gradient regularization is applied to reconstruct the super-resolution image. The proposed method is able to produce clear high-frequency texture details and maintain clean edges even under large scaling factors. Experimental results show that the proposed method performs well in visual effects and similarities. Furthermore, we test our algorithm in multi-texture images for robust evaluation. It is demonstrated that our algorithm is robust under complicated textures condition.
关键词： image reconstruction,super-resolution,fractional-order,blur kernel estimation
更新于2025-09-04 15:30:14
-
Confidence interval constraint based regularization framework for PET quantization
摘要： In this paper, a new generic regularized reconstruction framework based on confidence interval constraints for tomographic reconstruction is presented. As opposed to usual state-of-the-art regularization methods that try to minimize a cost function expressed as the sum of a data-fitting term and a regularization term weighted by a scalar parameter, the proposed algorithm is a two-step process. The first step concentrates on finding a set of images that relies on direct estimation of confidence intervals for each reconstructed value. Then, the second step uses confidence intervals as a constraint to choose the most appropriate candidate according to a regularization criterion. Two different constraints are proposed in this paper. The first one has the main advantage of strictly ensuring that the regularized solution will respect the interval-valued data-fitting constraint, thus preventing over-smoothing of the solution while offering interesting properties in terms of spatial and statistical bias/variance trade-off. Another regularization proposition based on the design of a smoother constraint also with appealing properties is proposed as an alternative. The competitiveness of the proposed framework is illustrated in comparison to other regularization schemes using analytical and GATE-based simulation and real PET acquisition.
关键词： confidence intervals,constrained regularization,Image reconstruction,total variation,positron emission tomography
更新于2025-09-04 15:30:14