Background Reconstruction Research Articles

Mura defect detection based on effective background reconstruction and contrast enhancement

Mura defect detection based on effective background reconstruction and contrast enhancement

An accurate reconstruction of CMB E-mode signal over large angular scales using prior information of CMB covariance matrix in ILC algorithm

ABSTRACT Recently, the internal-linear-combination (ILC) method was investigated extensively in the context of reconstruction of Cosmic Microwave Background (CMB) temperature anisotropy signal using observations obtained by WMAP and Planck satellite missions. In this article, we, for the first time, apply the ILC method to reconstruct the large-scale CMB E mode polarization signal, which could probe the ionization history, using simulated observations of 15 frequency CMB polarization maps of future generation Cosmic Origin Explorer (COrE) satellite mission. We find that the clean power spectra, from the usual ILC, are strongly biased due to non-zero CMB-foregrounds chance correlations. In order to address the issues of bias and errors we extend and improve the usual ILC method for CMB E-mode reconstruction by incorporating prior information of theoretical E-mode angular power spectrum while estimating the weights for linear combination of input maps (Sudevan & Saha 2018b). Using the E-mode covariance matrix effectively suppresses the CMB-foreground chance correlation power leading to an accurate reconstruction of cleaned CMB E-mode map and its angular power spectrum. We compare the performance of the usual ILC and the new method over large angular scales and show that the later produces significantly statistically improved results than the former. The new E-mode CMB angular power spectrum contains neither any significant negative bias at the low multipoles nor any positive foreground bias at relatively higher mutltipoles. The error estimates of the cleaned spectrum agree very well with the cosmic variance induced error.

A Novel Pixel-Wise Defect Inspection Method Based on Stable Background Reconstruction

In this article, an anomaly detection method based on background reconstruction is proposed to perform defect inspection on the texture surface of the industrial products. This method consists of two modules: 1) an autoencoder integrated with a generative adversarial network is utilized to reconstruct the textured background of the original image as a defect-free reference. Specifically, extra anomalous images are introduced and a mapping method of anomaly is given to improve the stability of reconstruction. 2) A U-net based inspection network is trained to perform pixel-wise analysis of the differences between the original and the reconstructed defect-free image. During these processes, only artificial synthesized defective images are utilized to train the model without any real defective samples. A series of experiments are conducted on several texture image data sets and the industrial production line. The experimental results reveal the effectiveness and versatility of the proposed method.

Moving target extraction and background reconstruction algorithm

It is difficult for the computer to distinguish the target from the background due to the long-time static of the target after moving. A new moving target detection and background reconstruction algorithm is proposed and is applied into the RGB video for the first time. Firstly, the proposed algorithm builds a model from the time dimension to extract the changed region. Then, it combines with the space dimension information to completely extract the moving target. The spatiotemporal correlation model is established to realize the construction of pure background. The experimental results show that the proposed algorithm can effectively reconstruct the background and the recognition rate of moving target is high.

Hypothenar Island flap: a safe and excellent choice for little finger defects

Objective The aim was to evaluate the functional and cosmetic results of hypothenar island flap for the reconstruction of little finger defects and to assess them by follow-up. Background Reconstruction of soft-tissue defects of little finger of the hand needs a single-stage procedure that can provide well-vascularized coverage at the earliest stage to give the best functional result. Patients and methods This follow-up study that included 19 patients (13 men and six women) was carried out in the Plastic and Reconstructive Surgery Departments of Menoufia University and Helmia Military Hospital from October 2017 to October 2019, with assessment by a follow-up period of 12 months and physiotherapy. All the patients studied had the hypothenar island flap for coverage of the defects. The defects were due to postblunt trauma in three cases, post-burn in three cases, and post-sharp trauma in 13 cases. The largest flap was 20 × 15 mm2. The donor site was directly closed in all cases. Results The overall results were satisfactory regarding the function and appearance without major complications according to patient and doctor satisfaction. Regarding flap course and survival, 14 (73.7%) flaps passed uneventfully, five (26.3%) flaps suffered from venous congestion, five congested flaps survived without total loss and only partial loss and thus we had 19 (100%) survived flaps. Conclusion Hypothenar island flap is a reliable and versatile source for coverage of soft-tissue defects of little finger of the hand.

Paradigmenwechsel – welches Paradigma?

Summary It would be too simple to attribute the current crisis in research on the Pentateuch to the demise of the Wellhausen paradigm (the documentary hypothesis) and the resulting diversity of methods. Rather, these are merely symptoms of a fundamental crisis in the source critical paradigm as a whole. The difficulties with the nature of sources, their literary scope and the criteria for their division are only part of the wider problem. The introduction of non-verifiable ‘redactors’ and ‘Fortschreibung’ (updating) to cope with anomalies of a hypothesis seems methodologically just as questionable as the reconstruction of the social-historical background of individual literary strata (‘pseudo historicism’). The increasingly complex diachronic overall models are lacking plausibility and credibility, especially since the reconstruction of several stages in the development of the text is practically impossible from an empirical perspective. The alternative to the classic paradigm of source criticism is a literary paradigm that returns to the starting point of the testimony of the biblical texts themselves. Of course, literary tensions and sources must not be ignored. However, systems of interpretation to explain inconsistencies should not be based on methods developed in the eighteenth and nineteenth centuries but on the ancient literary conventions of the Bible and its environment.

Discriminative Reconstruction Constrained Generative Adversarial Network for Hyperspectral Anomaly Detection

The rich and distinguishable spectral information in hyperspectral images (HSIs) makes it possible to capture anomalous samples [i.e., anomaly detection (AD)] that deviate from background samples. However, hyperspectral anomaly detection (HAD) faces various challenges due to high dimensionality, redundant information, and unlabeled and limited samples. To address these problems, this article proposes an unsupervised discriminative reconstruction constrained generative adversarial network for HAD (HADGAN). Our solution is mainly based on the assumption that the number of normal samples is much larger than the number of abnormal ones. The key contribution of this article is to learn a discriminative background reconstruction with anomaly targets being suppressed, which produces the initial detection image (i.e., the residual image between the original image and reconstructed image) with anomaly targets being highlighted and background samples being suppressed. To accomplish this goal, first, by using an autoencoder (AE) network and an adversarial latent discriminator, the latent feature layer learns normal background distribution and AE learns a background reconstruction as much as possible. Second, consistency enhanced representation and shrink constraints are added to the latent feature layer to ensure that anomaly samples are projected to similar positions as normal samples in the latent feature layer. Third, using an adversarial image feature corrector in the input space can guarantee the reliability of the generated samples. Finally, an energy-based spatial and distance-based spectral joint anomaly detector is applied in the residual map to generate the final detection map. Experiments conducted on several data sets over different scenes demonstrate its state-of-the-art performance.

Two-stage Reconstruction Using a Free Jejunum/Ileum Flap After Total Esophagectomy.

Reconstruction after esophagectomy is conventionally performed with a gastric conduit. However, in cases where a gastric conduit is unavailable, reconstructive procedures vary in terms of flap type, operative timing, and conduit route. Single-stage surgery is associated with a long operation time and high surgical stress, resulting in perioperative mortality. Recent advances in reconstructive microsurgery have made free intestinal flap transfer safe and reliable. Therefore, to overcome the shortcomings with previous methods, we performed 2-stage surgery involving free jejunum/ileum transfer for reconstruction after esophagectomy. From 2010 to 2018, 42 free jejunum/ileum flaps were transferred for reconstruction after esophagectomy in 41 patients. The diagnosis was esophageal cancer in 38 patients. All operations were performed in 2 stages. In most cases, total esophagectomy was performed in the first operation. The cervical stump of the esophagus was sutured to the cervical skin, creating an esophagostomy in the left neck. About 4 to 7 weeks after the first operation, the second operation was performed. The free jejunum/ileum flap was transferred through the subcutaneous route. Microvascular anastomosis was performed with the internal mammary artery and internal mammary vein, transverse cervical artery, internal and external jugular veins (internal jugular vein and EJV, respectively), and cephalic vein. The mean follow-up duration was 20 months. Free jejunum/ileum transfer was performed as the first operation in 4 cases and as the second operation in 38 cases. A free jejunal flap was used in 36 cases and free ileal flap was used in 6 cases. The recipient arteries were the internal mammary artery in 38 cases and transverse cervical artery in 4 cases. The recipient veins were the internal mammary vein in 15 cases, cephalic vein in 13 cases, EJV in 10 cases, and internal jugular vein in 10 cases. The flaps survived in all cases, except 1 case (41/42, 97.6%). The complications were anastomotic leakage of the flap in 9 cases, respiratory complications in 10 cases, and ileus in 2 cases. Perioperative mortality was not noted. Two-stage surgery using free jejunum/ileum flap transfer is a safe and reliable option for esophageal reconstruction in cases where gastric pull-up is unavailable.

3D-Printed Modular Hemipelvic Endoprosthetic Reconstruction Following Periacetabular Tumor Resection: Early Results of 80 Consecutive Cases.

Reconstruction with an endoprosthesis following pelvic tumor resection has increased over the years. However, the long-term results reflect a disappointing frequency of mechanical complications and failures. In an attempt to enhance implant fixation, an electron beam melting (EBM)-based modular hemipelvic endoprosthesis was introduced. Our aim was to explore the preliminary clinical outcomes for patients who have been managed with this prosthesis. We reviewed the records of 80 consecutive patients who had been managed at a single center between June 2015 and September 2017. Chondrosarcoma was the predominant diagnosis (31.3%). Osseous metastases were diagnosed in 16 patients (20.0%). The position of the reconstructed metallic acetabulum was measured on an anteroposterior pelvic radiograph. Bone ingrowth was evaluated in 2 samples harvested from patients with tumor recurrence. After a median duration of follow-up of 32.5 months (range, 9 to 52 months), no acetabular component instability was detected on radiographs. Histological sectioning of specimens harvested from 2 patients with tumor recurrence showed bone trabeculae extending toward the implant and bone ingrowth within the porous network. At the time of the latest follow-up, 59 patients (73.8%) were alive with no evidence of disease, 5 patients (6.3%) were alive with disease, and 16 patients (20.0%) had died of disease. Local recurrence occurred in 9 patients (11.3%). The mean Musculoskeletal Tumor Society score at the time of the latest follow-up was 83.9% (range, 43% to 100%). Complications occurred in 16 patients (20%), with wound dehiscence being the most common complication (8 patients; 10%). No aseptic loosening was found. Five patients (6.3%) had deep infection, and 2 patients (2.5%) had dislocation. The use of a 3-dimensional (3D)-printed modular hemipelvic endoprosthesis with a highly porous metal interface represents a potential choice as a pelvic endoprosthesis after internal hemipelvectomy for the treatment of a primary or metastatic tumor. These preliminary results demonstrate stable fixation with acceptable early functional and radiographic outcomes. Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.

An approach to evaluate myocardial perfusion defect assessment for projection-based DECT: A phantom study

IntroductionDual-energy CT (DECT) can improve the accuracy of myocardial perfusion CT with projection-based monochromatic (DECT-MCE) and quantification of myocardial iodine in material decomposition (DECT-MD) reconstructions. However, evaluation of multiple reconstructions is laborious and the optimal reconstruction to detect myocardial perfusion defects is unknown. MethodsLeft ventricular (LV) phantoms with artificial perfusion defects were scanned using DECT and single energy cardiac computed tomography angiography (SECT). Reconstructions of DECT-MCE at 40, 70, 100 and 140 keV, DECT-MD pairs of water, iodine, iron and fat, and SECT were evaluated using a 17-segment myocardial model. The diagnostic performance of each reconstruction was calculated on a per-segment basis and compared across DECT reconstructions. ResultsOver 34 phantoms with artificial perfusion defects were found in 64/578 (11%) of segments, the sensitivity of DECT-MCE at 40, 70, 100, and 140 keV was 100% (95% confidence interval (CI): 93–100), 100% (95% CI: 93–100), 71% (95% CI: 56–83), and 25% (95% CI: 14–40), respectively, with a significant decline between 70 keV and 100 keV (p < 0.001). The specificity of DECT-MCE was 100% at all energies (95% CI: 99–100). As a group, the DECT-MD iodine background reconstructions had significantly lower sensitivity than the remaining modes (2.1% [95% CI, 0.05–11.1], vs. 100% [95% CI, 92.6–100], p < 0.001). Specificity of all material pair modes remained 100%. ConclusionsUsing LV phantom models, the approach with the best sensitivity and specificity to assess myocardial perfusion defects with DECT are reconstructions of DECT-MCE at 40 or 70 KeV and DECT-MD without iodine background.

Attention-Driven Loss for Anomaly Detection in Video Surveillance

Recent video anomaly detection methods focus on reconstructing or predicting frames. Under this umbrella, the long-standing inter-class data-imbalance problem resorts to the imbalance between foreground and stationary background objects in video anomaly detection and this has been less investigated by existing solutions. Naively optimizing the reconstructing loss yields a biased optimization towards background reconstruction rather than the objects of interest in the foreground. To solve this, we proposed a simple yet effective solution, termed attention-driven loss to alleviate the foreground-background imbalance problem in anomaly detection. Specifically, we compute a single mask map that summarizes the frame evolution of moving foreground regions and suppresses the background in the training video clips. After that, we construct an attention map through the combination of the mask map and background to give different weights to the foreground and background region respectively. The proposed attention-driven loss is independent of backbone networks and can be easily augmented in most existing anomaly detection models. Augmented with attention-driven loss, the model is able to achieve AUC 86.0% on Avenue, 83.9% on Ped1, 96% on Ped2 datasets. Extensive experimental results and ablation studies further validate the effectiveness of our model.

3D reconstruction and validation of historical background for immersive VR applications and games: The case study of the Forum of Augustus in Rome

In the last decades, thanks to the success of the video games industry, the sector of technologies applied to cultural heritage has begun to envisage, in this domain, new possibilities for the dissemination of heritage and the study of the past through edutainment models. More recently, experimentation in the field of virtual archaeology has led to the development of virtual museums and interactive applications. Among these, the “serious game” segment – the application of interactive technologies to the cultural heritage domain – is rapidly growing, also including immersive VR technologies. Applied VR games and applications are characterized by a thorough historical background and a validated 3D reconstruction. Indeed, producing such products requires a tailored workflow and large effort in terms of time and professionals involved to guarantee such faithfulness. Drawing on our previous work in the field of virtual archaeology and referring to recent experiences related to the deployment of applied VR games on PlayStation VR ©, we describe and assess a workflow for the production of historically accurate 3D assets, targeting interactive, immersive VR products. The workflow is supported by the case study of the Forum of Augustus and different output applications, highlighting peculiarities and issues emerging from a multi and interdisciplinary approach.

Background Learning Based on Target Suppression Constraint for Hyperspectral Target Detection

Hyperspectral target detection is critical in both military and civilian applications. However, it is a challenging task due to the complexity of background and the limited samples of target in hyperspectral images (HSIs). In this article, we propose a novel background learning model, called background learning based on target suppression constraint to characterize high-dimensional spectral vectors. Considering insufficient target samples, the model is trained only on the background spectral samples to accurately learn the background distribution. Then the discrepancy between the reconstructed and original HSIs are examined to spot the targets. To obtain a background training dataset, coarse detection is carried out. However, it is quite difficult to retrieve pure background data. Thus, a target suppression constraint is imposed to reduce the impact of suspected target samples on background reconstruction. Experiments on six real HSIs demonstrate that the proposed framework significantly outperforms the current state-of-the-art detection methods and yields higher detection accuracy and lower false alarm rate.

Study of the effect of using antihemorrhagic drugs and the use of intraarticular drain after arthroscopic anterior cruciate ligament reconstruction on the outcome of the procedure

Background Reconstruction of the anterior cruciate ligament (ACL) is crucial for maintenance of the stability of the knee. ACL reconstruction is a bloody operation owing to the formation of transosseous tunnels. Blood might obscure the view during surgery and may collect postoperatively to impair the range of movement and slow down the rate of rehabilitation. The use of intraarticular (IA) drains might be helpful in that issue; however, the risk of introduction of infection to the joint might be increased. Hypothesis The purpose of this study was to determine whether the use of IA drains are necessary for improving outcomes after arthroscopic ACL reconstruction and also whether the use of preoperative antihemorrhagic drugs might preclude the use of IA drains. Patients and methods A total of 60 adult patients (60 knees) with documented ACL injury were eligible for arthroscopically assisted ACL reconstruction. The patients were randomly divided into two groups, each consisting of 30 patients. Both groups had anatomical single-bundle ACL reconstruction using the transaccessory femoral portal approach. The ACL graft substitute was hamstring tendon autograft in 25 cases and quadriceps tendon autograft in five cases. In the first group, an IA drain was fixed at the end of surgery, without preoperative antihemorrhagic drugs. The second group was given antihemorrhagic drugs IV (tranexamic acid 500 mg/5 ml, trade name Kapron) 1 h before surgery, and no drain was applied at the end of the operation. The patients were examined for the degree of knee swelling, pain as assessed by the visual analog scale, amount of blood collected, and range of movement. These parameters were reassessed after 2 weeks and at the end of the follow-up at 3 months postoperatively. Results Group 1 with the drain tended to have less pain and swelling in the immediate postoperative period, more wasting of the quadriceps at the end of follow-up, and comparatively lower mean range of motion (ROM) at the end of follow-up. On the contrary, group 2 without drain tended to have more swelling in the immediate postoperative period, which decreased dramatically over time to be significantly lower than that in patients in group 1 at a comparable stage of follow-up. Only four patients developed significant hemarthrosis that needed aspiration with a wide-bore needle in group 2. The final mean ROM is significantly better in this group. The rate of infection tended to be lower in group 2 without a drain, but with no statistically significant results. Conclusion The use of IA drain after arthroscopic ACL reconstruction decreases pain and swelling in the immediate postoperative period. However, patients among group 2 despite having more pain and swelling in the immediate postoperative period recovered more rapidly and more effectively, having a comparatively better ROM and less wasting of the quadriceps at the end of follow-up. The rate of infection tended to be lower in group 2 without a drain, but with no statistically significant results. The preoperative use of parenteral antihemorrhagic medication decreases the intraoperative bleeding and improves visualization during arthroscopic surgery; moreover, it may be beneficial in decreasing the postoperative hemarthrosis.

Unsupervised Disentanglement of Pose, Appearance and Background from Images and Videos.

Unsupervised landmark learning is the task of learning semantic keypoint-like representations without the use of expensive input keypoint annotations. A popular approach is to factorize an image into a pose and appearance data stream, then to reconstruct the image from the factorized components. The pose representation should capture a set of consistent and tightly localized landmarks in order to facilitate reconstruction of the input image. Ultimately, we wish for our learned landmarks to focus on the foreground object of interest. However, the reconstruction task of the entire image forces the model to allocate landmarks to model the background. Using a motion-based foreground assumption, this work explores the effects of factorizing the reconstruction task into separate foreground and background reconstructions in an unsupervised way, allowing the model to condition only the foreground reconstruction on the unsupervised landmarks. Our experiments demonstrate that the proposed factorization results in landmarks that are focused on the foreground object of interest when measured against ground-truth foreground masks. Furthermore, the rendered background quality is also improved as ill-suited landmarks are no longer forced to model this content. We demonstrate this improvement via improved image fidelity in a video-prediction task. Code is available at https://github.com/NVIDIA/UnsupervisedLandmarkLearning.

A Fast Surface Defect Detection Method Based on Background Reconstruction

In this paper, we propose an unsupervised background reconstruction method to detect defects on surfaces with unevenly distributed textures. An improved deep convolutional autoencoder is utilized to reconstruct the textured background of the original image as a defect-free reference. Specifically, a weighted loss function based on structural similarity (SSIM) is utilized to adapt to the unevenly distributed texture background and improve the reconstruction accuracy. Furthermore, combined with the reconstructed defect-free reference, a novel difference analysis method based on the discrete cosine transform (DCT) is given to accurately segment the defect regions from the original image. A series of experiments for the defect detection on mobile phone cover glass (MPCG) are conducted. The processing time for an image of 512 × 512 pixels is only 20 ms, which satisfies the requirement of online detection. The experimental results verify the effectiveness of the proposed method.

Split-thickness skin grafts obtained from adjacent hairy skin for reconstructing auricular concave surfaces after Mohs surgery.

Reconstruction of large auricular defects with full-thickness skin grafts (FTSG) is a commonly reported option, but less attention has focused on the advantages and indications of using split-thickness skin grafts (STSG) in the ear. We sought to report our experience using STSG for repair of defects located on the auricular concave surfaces, highlighting the utility of choosing the adjacent hairy skin as donor site. We performed a retrospective review of all Mohs micrographic defects on the auricular concave surfacesrepaired with STSG obtained from the adjacent hairy skin, between January 2017 and July 2018 at our institution. A total of 16 patients with defects on the auricular concavities resulting from removal of non-melanoma skin cancer were reconstructed with STSG taken from the adjacent hairy skin. Only one patient experienced partial graft failure and no other complications were observed after 6-month follow-up. Split-thickness skin grafts are suitable for reconstructing concave areas in the ear, providing good cosmetic results with a simple, cost-effective and easily reproducible technique. Choosing the adjacent hairy skin as a donor area shortens the operative and postoperative time, and allows the procedure to be performed in a single surgical field.

Second intention healing of nasal ala and dorsum defects in Asians

Background Reconstruction of defects of nasal ala and dorsum after surgical excision presents a substantial challenge to dermatologic surgeons. Second intention healing is a simple and extremely useful method to optimize cosmesis after skin cancer removal. Objectives This study reported the cosmetic outcomes after second intention healing of nasal ala and dorsum defects in Asians, and estimated the time to epithelialization and complete healing. Materials and methods Fifteen defects (<1 cm in diameter) of the nasal ala and dorsum in 10 patients were allowed to heal by secondary intention. Cosmetic results were evaluated and the time to epithelialization and complete healing were recorded. Results Cosmetic outcomes were good to excellent in 80% of the defects; defects of the dorsum showed poorer cosmetic results than defects of the ala. The wounds needed 5–17 days (mean 11.3; SD ± 4.18) to complete epithelialization and 10–24 days (mean 17.7; SD ± 4.85) to heal completely. Conclusions Second intention healing of small nasal ala and dorsum defects (<1 cm in diameter) in Asians produces satisfactory cosmetic results with a low complication rate.

Poezja, muzyka i sumienie. O Lukanie w Koronacji Poppei Giovanniego Francesco Busenellego i Claudio Monteverdiego

The article discusses the role of Lucano in the dramatic structure of The Coronation of Poppaea. After a deeper analysis, the poet, an episodic figure, turns out to be an important element in the reconstruction of the ideological background of the works by Busenello and Monteverdi. This is achieved by context reading of the opera libretto in two fundamental aspects – biographical and literary, based on two groups of textual sources related to Lucano – the first consisting of information about his life and work, and the second, which is his own works. Particular attention was paid to the analysis of Farsalia, a poem that shows numerous ideological affinities to Busenello’s libretto. A detailed analysis of literary contexts also reveals the self-thematic aspect of the scene with Lucano’s participation and the moral aspects of poetic creation that emerge in connection with it.

DeepCMB: Lensing reconstruction of the cosmic microwave background with deep neural networks

Next-generation cosmic microwave background (CMB) experiments will have lower noise and therefore increased sensitivity, enabling improved constraints on fundamental physics parameters such as the sum of neutrino masses and the tensor-to-scalar ratio r. Achieving competitive constraints on these parameters requires high signal-to-noise extraction of the projected gravitational potential from the CMB maps. Standard methods for reconstructing the lensing potential employ the quadratic estimator (QE). However, the QE performs suboptimally at the low noise levels expected in upcoming experiments. Other methods, like maximum likelihood estimators (MLE), are under active development. In this work, we demonstrate reconstruction of the CMB lensing potential with deep convolutional neural networks (CNN) - ie, a ResUNet. The network is trained and tested on simulated data, and otherwise has no physical parametrization related to the physical processes of the CMB and gravitational lensing. We show that, over a wide range of angular scales, ResUNets recover the input gravitational potential with a higher signal-to-noise ratio than the QE method, reaching levels comparable to analytic approximations of MLE methods. We demonstrate that the network outputs quantifiably different lensing maps when given input CMB maps generated with different cosmologies. We also show we can use the reconstructed lensing map for cosmological parameter estimation. This application of CNN provides a few innovations at the intersection of cosmology and machine learning. First, while training and regressing on images, we predict a continuous-variable field rather than discrete classes. Second, we are able to establish uncertainty measures for the network output that are analogous to standard methods. We expect this approach to excel in capturing hard-to-model non-Gaussian astrophysical foreground and noise contributions.