The morphological changes of the hippocampus in brain Magnetic Resonance Imaging (MRI) images are of great significance for the early screening of Alzheimer's disease. Currently, in clinical practice, the diagnosis of the hippocampus is achieved manually by doctors with experience. Because the hippocampus has the characteristics of small size, complex shape, and indistinct boundary with surrounding structures, manual segmentation, and grading of the hippocampus in brain MRI is time-consuming and labor-intensive, which is susceptible to errors because of human subjective judgment. To address that, this paper proposes a hippocampal MRI diagnosis algorithm based on Faster R-CNN and Mask R-CNN. The main contributions are 1) automatic identification of hippocampus in brain MRI by Faster R-CNN neural network, 2) precisely segmenting the hippocampus and judging the atrophy level through Mask R-CNN. Case studies are performed on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database and the medical records of the First Affiliated Hospital of Chongqing Medical University. Results indicate that the proposed method achieves a good segmentation effect on the hippocampus in the coronal MRI image of the brain and accurately grades the level of hippocampal atrophy, which can better assist doctors in diagnosing Alzheimer's disease.
The analytic solution of sparse signal reconstruction algorithm based on L1 regularization is a biased estimation, which leads to the underestimation of target intensity when applied to sparse SAR imaging, resulting in the bias effect and affecting the reconstruction accuracy. In this paper, we quantitatively analyze the bias effect in SAR imaging applications, and analyse the influence of target intensity, signal-to-noise ratio, intensity ratio of adjacent targets in the observation scene on the reconstruction bias. In order to suppress the bias effect and improve the reconstruction accuracy, we adopt a class of algorithms based on nonconvex penalty, and verify the performance of these algorithms using simulations and real data.
Polarimetric synthetic aperture radar (PolSAR) obtains polarimetric scattering of targets. The scattering properties are usually considered as invariant in azimuth. In some new SAR mode, such as wide-angle SAR and circular SAR (CSAR), targets are illuminated for longer time and look angle changes a lot. Moreover some targets have different physical shape in different look angle. Thus scattering properties can no longer be considered as invariant in azimuth. Variations across azimuth should be considered as useful information and are important parts of targets’ scattering properties. In this paper, polarimetric data are cut into subapertures in order to achieve scattering properties in different look angle. Target vector and coherency matrix are de- fined for multi-aperture situation. Polarimetric entropy for multi-aperture situation is then defined and named with multi-aperture poalrimetric entropy(MAPE). MAPE is calculated based on eigenvalue of multi-aperture coherency matrix. MAPE describes variations of scattering properties across subapertures. When MAPE is low, scattering properties change a lot across subapertures, which refers to anisotropic targets. When MAPE is high, there are few variations across subapertures, which refers to isotropic targets. Thus anisotropic targets and isotropic targets can be identified by MAPE. The effectiveness of MAPE is demonstrated on polarimetric CSAR(Pol-CSAR) data, acquired by the Institute of Electronics airborne CSAR system at P-band.
This paper presents an interferometric synthetic aperture radar (InSAR) imaging method based on L1 regularization reconstruction model for SAR complex-image and raw data via complex approximated message passing (CAMP) with joint reconstruction model. As an iterative recovery algorithm for L1 regularization, CAMP can not only obtain the sparse estimation of considered scene as other regularization recovery algorithms, but also a non-sparse solution with preserved background information, thus can be used to InSAR processing. The contributions of the proposed method are as follows. On the one hand, as multiple SAR complex images are strongly correlated, single-channel independent reconstruction via Lq regularization cannot preserve the interferometric phase information, while the proposed mixed norm-based L1 regularization joint reconstruction model via CAMP algorithm can ensure the preservation of interferometric phase information among multiple channels. On the other hand, the interferogram reconstructed by the proposed CAMP-based InSAR imaging with joint reconstruction model can improve the performance of noise reduction efficiently compared with conventional matched filtering (MF) results. Experiments carried out on simulated and real data confirmed the feasibility of the L1 regularization joint reconstruction model via CAMP for InSAR processing with preserved interferometric phase information and better noise reduction performance.
In this paper, we proposed an azimuth-range decouple-based L1 regularization method for wide ScanSAR imaging via extended chirp scaling (ECS) and applied it to the TerraSAR-X data to achieve large-scale sparse reconstruction. Compared with ECS, the conventional ScanSAR imaging algorithm based on matched filtering, the proposed method can improve the synthetic aperture radar image performance with full-sampling raw data for not only sparse but also nonsparse surveillance regions. It can also achieve high-resolution imaging for sparse considered scenes efficiently using down-sampling raw data. Compared with a typical L1 regularization imaging approach, which requires transfer of the two-dimensional (2-D) echo data into a vector and reconstruction of the scene via 2-D matrix operation, our proposed method has less computational cost and hence makes the large-scale regularization reconstruction of considered area become possible. The experimental results via real data validate the effectiveness of the proposed method.
In this paper, we develop a group sparsity based wide angle synthetic aperture radar (WASAR) imaging model and propose a novel algorithm called backprojection based group complex approximate message passing (GCAMP-BP) to recover the anisotropic scene. Compare to conventional backprojection based complex approximate message passing (CAMP-BP) algorithm for the recovery of isotropic scene, the proposed method accommodates aspect dependent scattering behavior better and can produce better imagery. Simulated and experimental results are presented to demonstrate the validity of the proposed algorithm.
The elevation image quality of tomographic synthetic aperture radar (TomoSAR) data depends mainly on the elevation aperture size, number of baselines, and baseline distribution. In TomoSAR, due to the restricted number of baselines with irregular distributions, the elevation imaging quality is always unacceptable using the conventional spectral analysis approach. Therefore, for a given limited number of irregular baselines, the completion of data for the unobserved virtual uniform baseline distribution should be addressed to improve the spectral analysis-based TomoSAR reconstruction quality. We propose an Lq(0
KEYWORDS: 3D image processing, Synthetic aperture radar, Reconstruction algorithms, Radar imaging, Stereoscopy, Signal to noise ratio, Imaging systems, 3D acquisition, 3D modeling, Antennas
We propose an imaging algorithm for downward-looking sparse linear array three-dimensional synthetic aperture radar (DLSLA 3-D SAR) in the circumstance of cross-track sparse and nonuniform array configuration. Considering the off-grid effect and the resolution improvement, the algorithm combines pseudo-polar formatting algorithm, reweighed atomic norm minimization (RANM), and a parametric relaxation-based cyclic approach (RELAX) to improve the imaging performance with a reduced number of array antennas. RANM is employed in the cross-track imaging after pseudo-polar formatting the DLSLA 3-D SAR echo signal, then the reconstructed results are refined by RELAX. By taking advantage of the reweighted scheme, RANM can improve the resolution of the atomic norm minimization, and outperforms discretized compressive sensing schemes that suffer from off-grid effect. The simulated and real data experiments of DLSLA 3-D SAR verify the performance of the proposed algorithm.
Recent theory of compressed sensing (CS) has been widely used in many application areas. In this paper, we mainly
concentrate on the CS in radar and analyze the distinguishing ability of CS radar image based on information theory
model. The information content contained in the CS radar echoes is analyzed by simplifying the information
transmission channel as a parallel Gaussian channel, and the relationship among the signal-to-noise ratio (SNR) of the
echo signal, the number of required samples, the length of the sparse targets and the distinguishing level of the radar
image is gotten. Based on this result, we introduced the distinguishing ability of the CS radar image and some of its
properties are also gotten. Real IECAS advanced scanning two-dimensional railway observation (ASTRO) data
experiment demonstrates our conclusions.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.