Projection Quality Research Articles

Camera-view supervision for bird's-eye-view semantic segmentation

Bird's-eye-view Semantic Segmentation (BEVSS) is a powerful and crucial component of planning and control systems in many autonomous vehicles. Current methods rely on end-to-end learning to train models, leading to indirectly supervised and inaccurate camera-to-BEV projections. We propose a novel method of supervising feature extraction with camera-view depth and segmentation information, which improves the quality of feature extraction and projection in the BEVSS pipeline. Our model, evaluated on the nuScenes dataset, shows a 3.8% improvement in Intersection-over-Union (IoU) for vehicle segmentation and a 30-fold reduction in depth error compared to baselines, while maintaining competitive inference times of 32 FPS. This method offers more accurate and reliable BEVSS for real-time autonomous driving systems. The codes and implementation details and code can be found at https://github.com/bluffish/sucam.

Controlling the scatterplot shapes of 2D and 3D multidimensional projections

Multidimensional projections are effective techniques for depicting high-dimensional data. The point patterns created by such techniques, or a technique’s visual signature, depend — apart from the data themselves — on the technique design and its parameter settings. Controlling such visual signatures — something that only few projections allow — can bring additional freedom for generating insightful depictions of the data. We present a novel projection technique — ShaRP — that allows explicit control on such visual signatures in terms of shapes of similar-value point clusters (settable to rectangles, triangles, ellipses, and convex polygons) and the projection space (2D or 3D Euclidean or S2). We show that ShaRP scales computationally well with dimensionality and dataset size, provides its signature-control by a small set of parameters, allows trading off projection quality to signature enforcement, and can be used to generate decision maps to explore the behavior of trained machine-learning classifiers.

Organized Optimization Integration Validation Model for Internet of Things (IoT)-Based Real-Time Applications

Emerging technology like the Internet of Things (IoT) has great potential for use in real time in many areas, including healthcare, agriculture, logistics, manufacturing, and environmental surveillance. Many obstacles exist alongside the most popular IoT applications and services. The quality of representation, modeling, and resource projection is enhanced through interactive devices/interfaces when IoT is integrated with real-time applications. The architecture has become the most significant obstacle due to the absence of standards for IoT technology. Essential considerations while building IoT architecture include safety, capacity, privacy, data processing, variation, and resource management. High levels of complexity minimization necessitate active application pursuits with variable execution times and resource management demands. This article introduces the Organized Optimization Integration Validation Model (O2IVM) to address these issues. This model exploits k-means clustering to identify complexities over different IoT application integrations. The harmonized service levels are grouped as a single entity to prevent additional complexity demands. In this clustering, the centroids avoid lags of validation due to non-optimized classifications. Organized integration cases are managed using centroid deviation knowledge to reduce complexity lags. This clustering balances integration levels, non-complex processing, and time-lagging integrations from different real-time levels. Therefore, the cluster is dissolved and reformed for further integration-level improvements. The volatile (non-clustered/grouped) integrations are utilized in the consecutive centroid changes for learning. The proposed model’s performance is validated using the metrics of execution time, complexity, and time lag.

ALVEOLAR DISTRACTION OSTEOGENESIS AND IMPLANTOLOGY

Congenital or acquired alveolar abnormalities are the indications for alveolar ridge augmentation. Periodontal disease, traumatic avulsions, post-extraction abnormalities, and/or protracted denture use with ensuing disuse atrophy can all contribute to acquiredalveolar bone loss. The horizontal dimension has suffered the most loss in the majority of these instances. The loss of the buccal bone plate following a traumatised tooth avulsion is a common scenario that results in a horizontal defect. The goal of current implant dentistry standards is to produce natural-looking prosthetic restorations with the best possible functional and aesthetic results. In order to attain gold-standard outcomes, a number of factors have been proposed, including sufficient bone height, width, and anteroposterior projection sufficient amount and quality of soft tissue preservation of the buccal sulcus and sufficient papillae and gingival contour. In the modern practice of dentistry, the preservation and rebuilding of the alveolar bone and surrounding soft tissues for the implantation of dental implants has become essential. An alternate technique for repairing alveolar atrophy is alveolar widening by distraction osteogenesis (DO), which is comparable to alveolar split grafting but does not require the graft. The alveolar bone and associated mucosa have undergone a durable, aesthetically pleasing restoration thanks to the combination of vertical DO and osseointegration.

Thermographic procedure for the assessment of Resistance Projection Welds (RPW): Investigating parameters and mechanical performances

This study presents a non-destructive testing (NDT) thermographic procedure for assessing the quality and mechanical strength of Resistance Projection Welded (RPW) joints with rectangular embossments. We analysed twelve RPW joints by systematically varying process parameters based on a factorial design. These joints underwent flash thermography followed by mechanical tests to evaluate the maximum breaking force (Fmax). Significant statistical correlations between process parameters (time and force) and Fmax were established. Furthermore, we found a correlation (p-value 0.86) between the optically measured fused region and Fmax. Subsequently, we developed a pulsed phase thermography-based procedure for non-destructively measuring the fused region, resulting in an average difference of approximately 4 % compared to optical measurements. An empirical linear relationship was derived to correlate the welded area obtained by thermal data with Fmax, enabling the estimation of mechanical joint strength through non-destructive pulsed thermography. This research offers a promising approach for assessing the mechanical integrity of RPW joints using thermal imaging techniques.

MF+C: Linking suboptimal projections to detail on handheld devices

Mobile Focus + Context (mF+C) involves using a handheld device as a focus screen for content on an immersive display or mobile projector. In this work we examine how using a focus device can mitigate poor context image quality due to environmental factors. In an exploratory study we compare three techniques for linking focus and context: lens-focus (Lens), where the device works as a mobile lens held parallel to and in front of the context, center-focus (Centered), where the user holds the device in the center of the projection and pans both context and focus by swiping, and marker-focus (Marker), where the focus region is highlighted on the context, and the user pans the focus by swiping. Participants performed searching, tracing, and detail acquisition tasks with maps and electronics diagrams under a range of simulated projection conditions. All techniques were able to mitigate poor projection quality and performed comparably in time and precision, but the effectiveness of each technique was impacted by task type. There was lower variation in time between participants using Lens for tracing tasks versus the other techniques, but wider variation for searching tasks. Tasks completed using sub-optimal projections involved more time spent looking at the context image than tasks with clear projections, however this difference is less pronounced for the Lens technique. We propose a hybrid Lens-Marker approach for mobile Focus+Context applications in dynamic environments.

Improving reduced-order models through nonlinear decoding of projection-dependent outputs

A fundamental hindrance to building data-driven reduced-order models (ROMs) is the poor topological quality of a low-dimensional data projection. This includes behavior such as overlapping, twisting, or large curvatures or uneven data density that can generate nonuniqueness and steep gradients in quantities of interest (QoIs). Here, we employ an encoder-decoder neural network architecture for dimensionality reduction. We find that nonlinear decoding of projection-dependent QoIs, when embedded in a dimensionality reduction technique, promotes improved low-dimensional representations of complex multiscale and multiphysics datasets. When data projection (encoding) is affected by forcing accurate nonlinear reconstruction of the QoIs (decoding), we minimize nonuniqueness and gradients in representing QoIs on a projection. This in turn leads to enhanced predictive accuracy of a ROM. Our findings are relevant to a variety of disciplines that develop data-driven ROMs of dynamical systems such as reacting flows, plasma physics, atmospheric physics, or computational neuroscience.

Automated F18-FDG PET/CT image quality assessment using deep neural networks on a latest 6-ring digital detector system

To evaluate whether a machine learning classifier can evaluate image quality of maximum intensity projection (MIP) images from F18-FDG-PET scans. A total of 400 MIP images from F18-FDG-PET with simulated decreasing acquisition time (120 s, 90 s, 60 s, 30 s and 15 s per bed-position) using block sequential regularized expectation maximization (BSREM) with a beta-value of 450 and 600 were created. A machine learning classifier was fed with 283 images rated “sufficient image quality” and 117 images rated “insufficient image quality”. The classification performance of the machine learning classifier was assessed by calculating sensitivity, specificity, and area under the receiver operating characteristics curve (AUC) using reader-based classification as the target. Classification performance of the machine learning classifier was AUC 0.978 for BSREM beta 450 and 0.967 for BSREM beta 600. The algorithm showed a sensitivity of 89% and 94% and a specificity of 94% and 94% for the reconstruction BSREM 450 and 600, respectively. Automated assessment of image quality from F18-FDG-PET images using a machine learning classifier provides equivalent performance to manual assessment by experienced radiologists.

Безопасная эксплуатация зданий и сооружений на транспорте

Purpose: To analyze the organization of safe operation of transport infrastructure facilities, their maintenance, routine repairs and working control in terms of building constructions. To establish the imperfections of regulatory documents for building and structure projection, construction and exploitation, to develop recommendations on quality improvement for facility safe operation organization. Methods: Analysis of regulatory documents (State Standards, Codes of Rules, Technical Regulations, Federal Laws, etc.), the results of technical inspection of constructional structures of buildings and facilities, the quality assessment of construction and exploitation of facilities including transport infrastructural ones. Results: Overview and analysis of the organization of safe operation of transport infrastructure facilities on all stages of lifecycle are given. The reliability and operability of constructional structures of buildings and facilities on transport depends on projection quality and the presence in specified documentation of full value section on facility exploitation, compliance of construction-installation works with specified documentation, timeliness, the presence there of the completeness and quality of working control, maintenance and routine repairs of a facility. Acting regulatory documents require particular practical recommendations or references on terminology regulations, additions and clarifications. The specificities of maintenance, routine repair and working control for constructional structures of buildings and facilities, the issues of their conformity and condition assessment as well as of reliability support in accordance with acting legislation are considered. Practical significance: "Instructions for Safe Exploitation of Foundations and Constructional Structures of Infrastructural Objects of St. Petersburg Subway" has been developed which has been introduced into the practical activities of Tunnel Construction Service for State Unitary Enterprise "St. Petersburg Subway", imperfections of regulatory documents at projection, construction and exploitation stage have been identified, recommendations on improving the quality to organize safe exploitation of buildings and constructions during the development of regulatory documents have been proposed.

Cultura organizacional según el modelo del cubo de la doble S

Culture has an important effect on management and knowing the type of culture is valuable information for the management of any type of organization. The present investigation is carried out with the objective of knowing what type of organizational culture a faculty has in a Peruvian public university, applying the double S cube model, which supposes two variables; sociability and solidarity (double S). Applying the corresponding questionnaire, it is obtained that the faculty mainly has a fragmented culture in 03 of its academic departments and only in 01 department does it have a communal culture, the latter due to its significantly higher level of solidarity with respect to what the other 03 departments had. Among the consequences for management, it is pointed out that, the type of culture fragmented due to its characteristics will hinder some policies to improve the quality of teaching, research and social projection.

Speckle Noise-Free Interconnective Holographic Projection

Generally, speckle noise is regarded as unavoidable in holographic projection, and it results from unexpected high spatial frequency components of diffracted light at the sharp edge of pixel openings in a spatial light modulator. Speckle noise typically reduces image contrast and degrades the image quality of the holographic projection. In this study, we propose a novel holographic optical interconnection method free of speckle noise in holographic projection. This optical interconnection is achieved by using a holographic optical element (HOE). The HOE is designed to reconstruct Gaussian beams with low divergence. These Gaussian beams become points which form target images at desired depths. Since the Gaussian beam from the HOE does not share the same position with other Gaussian beams, there is no interference at the projection image. Therefore, the image is composed of the points from the Gaussian beams and there is no reason for unexpected high spatial frequency noise to appear on the image. In this paper, we fabricate the HOE, produced with our specially manufactured hologram printer, where the directions of two Gaussian beams with low divergence are controlled by goniometers. We experimentally demonstrated a speckle noise-free interconnective holographic projection. Two images are successfully formed at different depths by optically connecting two points in pairs.

Phase Target-Based Calibration of Projector Radial Chromatic Aberration for Color Fringe 3D Measurement Systems.

The camera and projector are indispensable hardware parts of a color fringe projection 3D measurement system. Chromatic aberration between different color channels of the projector and camera has an impact on the measurement accuracy of the color fringe projection 3D profile measurement. There are many studies on camera calibration, but the chromatic aberration of the projector remains a question deserving of further investigation. In view of the complex system architecture and theoretical derivation of the traditional projector radial chromatic aberration method, a phase target based on projector radial chromatic aberration measurement and the correction method are proposed in this paper. This method uses a liquid crystal display with a holographic projection film as the phase target. The liquid crystal display sequentially displays red, green, and blue horizontal and vertical sinusoidal fringe images. The projector projects red, green, and blue horizontal and vertical sinusoidal fringe images to the phase target in turn, and calculates the absolute phases of the display fringes and reflection fringes, respectively. Taking the green channel as the reference channel, a phase coordinate system is established based on the phases of the vertical and horizontal directions displayed on the display screen, using the phase of the reflection fringes on the display screen as the ideal phase value of the phase point. Then, the phase coordinate system of the red and blue channels is transferred to the green phase coordinate system to calculate the chromatic aberration of the red-green channels and the blue-green channels, and pre-compensation is conducted. Experimental results prove that this method can measure and calibrate the radial chromatic aberration of the projector without being affected by the image quality of the camera. The correction effect of this method is that the maximum chromatic aberration of the red-green channel decreases from 1.9591/pixel to 0.5759/pixel, and the average chromatic aberration decreases from 0.2555/pixel to 0.1865/pixel. In addition, blue-green channel maximum chromatic aberration decreased from 1.8906/pixel to 0.5938/pixel, and the average chromatic aberration decreased from 0.2347/pixel to 0.1907/pixel. This method can improve the projection quality for fringe projection 3D profile measurement technology.

3D-printed model for preclinical training in oral and maxillofacial radiology

<h3>Objective</h3> 3D printing is experiencing significant growth in the teaching and learning process. Therefore, this study aimed to present a 3D-printed skull model for preclinical intraoral radiographic practice. <h3>Study Design</h3> Two 3D-printed mannequins were created. The first was created using an STL file of a skull and was edited using two 3D modeling softwares (Meshmixer and Netfabb). The second mannequin was designed directly from segmentation of a patient's CBCT data and then converted into an STL file. Both mannequins were printed using fused deposition modeling (FDM) technology and polylactic acid (PLA) filament. The printed skull bones were attached, and the mandible was articulated to the articular fossa of the temporal bone. The teeth were inserted into the alveoli. Intraoral radiographs of both mannequins were acquired using a digital sensor (Carestream RVG 5100). <h3>Results</h3> Both 3D-printed mannequins showed satisfactory radiographic appearance, allowing geometric representation of each intraoral radiographic projection, regardless of STL file origin. Anatomical structures, such as the periodontal ligament space, zygomatic process of the maxilla, and intermaxillary suture, were represented. The material cost of the printed prototype was $34.00. <h3>Conclusions</h3> The use of 3D-printed models is presented as an alternative to artificial commercial phantoms for the preclinical training of intraoral radiographic techniques, combining the radiographic projection's quality, the possibility of model manipulation, and an affordable price. <b>Statement of Ethical Review</b> Ethical review was sought and study was exempted from ethical review

A Motion Artifact Correction Algorithm for Cone-Beam CT in Patients with Hepatic Malignancies Treated with Transarterial Chemoembolization

PurposeTo evaluate the effect of a motion artifact correction algorithm (MACA) on cone-beam computed tomography (CT) during transarterial chemoembolization for hepatic malignancies. Materials and MethodsFrom June 2020 to March 2021, 42 patients with mild-to-severe motion artifacts detected using single cone-beam CT scans were evaluated retrospectively. The image quality of native and motion-corrected data was compared. The maximum intensity, sharpness, and full width at half maximum (FWHM) of 5 segmental hepatic arteries were quantitatively measured. The overall quality of maximum intensity projection (MIP) images, conspicuity of tumor-supplying arteries, and need for selective angiography to ascertain the vascular anatomy were qualitatively evaluated by multiple readers. Paired t and Wilcoxon signed rank tests were used to compare the parameters. ResultsThe mean maximum intensity and sharpness increased from 2,792.01 HU ± 451.36 to 3,148.40 HU ± 594.46 and from 0.31 ± 0.02/mm to 0.34 ± 0.02/mm, respectively, using the MACA (both P < .001). The MACA decreased the mean FWHM from 2.02 mm ± 0.27 to 1.78 mm ± 0.26 (P < .001). The overall quality of the MIP images and the conspicuity of the tumor-supplying artery were enhanced from 2.5 to 3.0 points and from 3.0 to 4.0 points, respectively (both P < .001). Selective angiography was expected to be omitted in 7 cases (16.7%, 7/42) after using the MACA. ConclusionsThe MACA significantly improved both quantitative and qualitative image quality of cone-beam CT in selected patients with motion artifacts during transarterial chemoembolization for hepatic malignancies.

Production of Head-Up Display windshield and its relation with the image quality

Previously, car windshields served as a barrier against the wind or to prevent the driver from flying outside the vehicle in a collision. The windshield is therefore laminated and consists of two glasses and an intermediate layer consisting of the so-called polyvinyl butyral foil. The latest trend in car windshields is to consider the windshield as a display and project key information onto the windshield. Such glasses are called head-up displays. Various information such as car speed or route can be projected. The driver thus constantly maintains contact with the road, which increases driving safety. Head-Up Display quality is key to display information correctly without negative effects. Such a negative effect is if the displayed information is duplicated, which is called ghosting or double-image. This effect is undesirable and ranks first among the defects we detect in Head-Up Displays. The aim of this article is to compare different production technologies with a focus on the quality of HUD projection and on the double-image occurrence. The production technology enabling optimal HUD image quality is discussed as well. The article deals with a summary of HUD production technologies, focusing on the most common defect that occurs in HUD windshields and which reduce the overall quality of the product. The article evaluates and summarizes the key findings from the perspective of the end customer, but also in terms of the production process. For the driver, as an end customer, it is a source of information according to which he can orient himself in the offer of car equipment. It can be useful for production or development people in terms of processability and process costs. At the same time, the benefits and negatives of using windshields with a Head-Up Display are discussed.

Semi-automatic data annotation based on feature-space projection and local quality metrics: An application to cerebral emboli characterization.

We propose a semi-supervised learning approach to annotate a dataset with reduced requirements for manual annotation and with controlled annotation error. The method is based on feature-space projection and label propagation using local quality metrics. First, an auto-encoder extracts the features of the samples in an unsupervised manner. Then, the extracted features are projected by a t-distributed stochastic neighbor embedding algorithm into a two-dimensional (2D) space. A selection of the best 2D projection is introduced based on the silhouette score. The expert annotator uses the obtained 2D representation to manually label samples. Finally, the labels of the labeled samples are propagated to the unlabeled samples using a K-nearest neighbor strategy and local quality metrics. We compare our method against semi-supervised optimum-path forest and K-nearest neighbor label propagation (without considering local quality metrics). Our method achieves state-of-the-art results on three different datasets by labeling more than 96% of the samples with an annotation error from 7% to 17%. Additionally, our method allows to control the trade-off between annotation error and number of labeled samples. Moreover, we combine our method with robust loss functions to compensate for the label noise introduced by automatic label propagation. Our method allows to achieve similar, and even better, classification performances compared to those obtained using a fully manually labeled dataset, with up to 6% in terms of classification accuracy.

Novel Projection Schemes for Graph-Based Light Field Coding.

In light field compression, graph-based coding is powerful to exploit signal redundancy along irregular shapes and obtains good energy compaction. However, apart from high time complexity to process high dimensional graphs, their graph construction method is highly sensitive to the accuracy of disparity information between viewpoints. In real-world light field or synthetic light field generated by computer software, the use of disparity information for super-rays projection might suffer from inaccuracy due to vignetting effect and large disparity between views in the two types of light fields, respectively. This paper introduces two novel projection schemes resulting in less error in disparity information, in which one projection scheme can also significantly reduce computation time for both encoder and decoder. Experimental results show projection quality of super-pixels across views can be considerably enhanced using the proposals, along with rate-distortion performance when compared against original projection scheme and HEVC-based or JPEG Pleno-based coding approaches.

P‐63: Design and Fabrication of Wide‐Viewing‐Angle Ambient Light Rejection Front Projection Screen

The image quality of projection displays is decreased by ambient light. This paper proposed a screen that composed of a microlens array film with printed black ink stripes on back side and a matte‐white‐reflection film to reduce the influence of ambient light in the front projection system. The ambient light was guided to light absorptive black ink by micro‐lens and thus enhance the contract ratio of the image in the bright environment. A proposed screen performed a better than 6.5 times improvement in image contrast ratio without sacrificing viewing angle compared to a matte‐white screen under ambient light condition was demonstrated. The proposed projection screen works well in short‐to‐standard throw projectors and good for home or public applications.

Three-Dimensional printed model for preclinical training in oral radiology.

3D printing is experiencing significant growth in the teaching and learning process. This study aims to present a 3D printed skull model for preclinical intraoral radiographic practice. Two 3D printed mannequins were created. One mannequin used an STL file of a skull that was edited using two 3D modelling software (Meshmixer and Netfabb). The second mannequin was designed directly from a patient's segmented CBCT data and then converted into an STL file. Both mannequins were printed using fused deposition modelling (FDM) technology and polylactic acid (PLA) filament; teeth for the second mannequin were also printed using digital light processing (DLP). The printed skull bones were attached, the mandible was articulated to the articular fossa of the temporal bone, and the teeth were inserted into the alveoli. Intraoral radiographs of both mannequins were taken using a digital sensor (RVG 5100, Carestream). Both 3D printed mannequins showed satisfactory radiographic appearance, allowing geometric representation of each intraoral radiographic projection, regardless of STL file origin. Anatomical structures, such as the periodontal ligament space, zygomatic process of the maxilla and intermaxillary suture, were represented. The material cost of the first and second printed prototype was $34.00 and $39.00, respectively. The use of 3D printed models is presented as an alternative to artificial commercial phantoms for the preclinical training of intraoral radiographic techniques through the combined benefits of superior radiographic projection quality, the possibility of model manipulation and an affordable price.

Single-Shot Intrinsic Calibration for Autonomous Driving Applications.

In this paper, we present a first-of-its-kind method to determine clear and repeatable guidelines for single-shot camera intrinsic calibration using multiple checkerboards. With the help of a simulator, we found the position and rotation intervals that allow optimal corner detector performance. With these intervals defined, we generated thousands of multiple checkerboard poses and evaluated them using ground truth values, in order to obtain configurations that lead to accurate camera intrinsic parameters. We used these results to define guidelines to create multiple checkerboard setups. We tested and verified the robustness of the guidelines in the simulator, and additionally in the real world with cameras with different focal lengths and distortion profiles, which help generalize our findings. Finally, we used a 3D LiDAR (Light Detection and Ranging) to project and confirm the quality of the intrinsic parameters projection. We found it possible to obtain accurate intrinsic parameters for 3D applications, with at least seven checkerboard setups in a single image that follow our positioning guidelines.