Error Propagation Research Articles

It's about time: moving away from statistical analysis of ecotoxicity data.

Environmental risk assessment of chemicals (ERA) relies on single-species laboratory testing to establish the toxic properties of a compound. However, ERA is not concerned with toxicity under laboratory conditions: it needs to assess the impacts of the compound in the real world. Data-driven statistical analyses (e.g., hypothesis testing and interpolation) are the common approaches for analysing toxicity data, but such approaches are the wrong tool for the job at hand. ERA does not need a statistical description of the effects in the toxicity test (at the end of the standardised test duration), it needs to extrapolate from the laboratory test to longer and time-varying exposure. Such extrapolation requires mechanistic process models, providing a simplified representation of the mechanisms underlying toxicity. Any useful model for the toxicity process should explicitly consider both dose (e.g., exposure concentration) and time. In the history of effects analysis for ERA, the factor of time does not get as much attention as the dose, hence common use of the term 'dose-response analysis'. However, this is a historical oversight: time is a crucial factor for understanding toxicity and thereby essential for meaningful extrapolation from laboratory to field. Mechanistic models for ecotoxicity, considering both dose and time, have been around for quite some time and are classified as toxicokinetic-toxicodynamic (TKTD) models. TKTD models are starting to find their way into pesticide ERA in Europe, next to the classical statistical approaches. In this opinion paper, I argue that it is about time to leave statistical analysis of toxicity data behind us. Statistics remains important for ERA's effects assessment, but its role lies in the definition of appropriate 'error models', explaining the deviations between model output and observations, which is essential for parameter estimation, uncertainty quantification, and error propagation. The 'process model', essential for extrapolation, firmly belongs to TKTD modelling.

Practical error prediction in UAV imagery-based 3D reconstruction: assessing the impact of image quality factors

ABSTRACT Over the past decade, Structure-from-Motion (SfM) and Multi-View Stereo (MVS) techniques have been highly effective in generating high-quality 3D point clouds, especially when integrated with Unmanned Aerial Vehicles (UAVs). However, accurately predicting errors in these point clouds remains challenging. This study introduces a predictive model for quantifying errors in point clouds generated using SfM-MVS, based on 2D image error propagation. We analysed the impact of four key image quality factors – exposure, resolution, blur (including motion and out-of-focus blur), and noise – on 2D image errors. These factors are determined using nine practical parameters belonging to camera settings (ISO, shutter speed, F-number, and pixel resolution), UAV flight methods (speed and distance), camera specifications (sensor size and focal length), and illumination conditions. To account for variations in SfM-MVS output quality due to software and camera model differences, we introduced two experimentally calibratable constants: the maximum template size and the noise sensitivity coefficient. We validated our error prediction model by comparing the predicted errors with observations from a reference indoor target, utilizing a total of 258 image sets acquired with varying parameter combinations. The model demonstrated high predictive accuracy, achieving an R2 value of 0.84. This confirms the effectiveness and feasibility of our model in accurately predicting point cloud errors using the nine parameters.

A Trajectory Prediction Method for Reentry Glide Vehicles via Adaptive Cost Function

This paper proposes a trajectory prediction method via the adaptive cost function to address the difficulties in inferring the attack intention and maneuver mode, as well as the accumulation of prediction error during the trajectory prediction of reentry glide vehicles. Firstly, the vehicle guidance task is divided into two distinct categories: conventional guidance and no-fly zone avoidance guidance. A task-matched time-varying parameter prediction model set is then constructed. Secondly, taking into account the maneuverability, guidance intent, and battlefield situation of the vehicle, an adaptive intent cost function adapted to the guidance task is proposed, which avoids the estimation failure problem caused by manually setting cost coefficients in traditional methods. Finally, long-term trajectory prediction of vehicles is achieved using Bayesian theory to infer the attack intent and parametric model with the maximum a posteriori probability. The results of the simulations demonstrate that the proposed prediction method is capable of accurately inferring the vehicle’s attack intention and parameter model, and of effectively reducing the accumulation of prediction errors and the time required for the algorithmic process compared to existing methods.

Deep learning-based 6-DoF visual relocalization assisted simultaneous localization and mapping (SLAM)

Purpose Visual simultaneous localization and mapping (SLAM) methods suffer from accumulated errors, especially in challenging environments without loop closure. By constructing lightweight offline maps and using deep learning (DL)-based technology in the two stages, i.e. image retrieval and feature matching, the goal is to reconstruct the six-degree-of-freedom (6-DoF) relationship between SLAM sequences and map sequences. This study aims to propose a comprehensive coarse-to-fine 6-DoF long-term visual relocalization assisted SLAM method specifically designed for challenging environments, aiming to achieve more accurate pose estimation. Design/methodology/approach First, image global feature matching and patch-level global feature matching are conducted to achieve optimal frame-to-frame matching. Second, a DL network is introduced to extract and match features between the most similar frames, enabling point-to-point motion estimation. Finally, a fast pose graph optimization method is proposed to achieve real-time optimization of the pose in the SLAM sequence. Findings The proposed method has been successfully validated on the real-world FinnForest Dataset and UZH-FPV Drone Racing Dataset. The accuracy of the proposed method is evaluated using absolute positional error and absolute rotational error. Experimental results show that in most cases, there are significant improvements in the root mean square error and the standard deviation of the error in pose estimation, and it performs better than loop closure in terms of accuracy. This indicates that the method has strong generalizability and robustness. Originality/value The main contribution of this study is the proposal of a complete DL-based coarse-to-fine 6-DoF long-term visual relocalization method to assist vSLAM, which demonstrates enhanced robustness and generalizability and can eliminate cumulative errors in pose estimation under challenging environments.

An efficient deep unrolling network for sparse-view CT reconstruction via alternating optimization of dense-view sinograms and images

Objective. Recently, there have been many advancements in deep unrolling methods for sparse-view computed tomography (SVCT) reconstruction. These methods combine model-based and deep learning-based reconstruction techniques, improving the interpretability and achieving significant results. However, they are often computationally expensive, particularly for clinical raw projection data with large sizes. This study aims to address this issue while maintaining the quality of the reconstructed image.Approach. The SVCT reconstruction task is decomposed into two subproblems using the proximal gradient method: optimizing dense-view sinograms and optimizing images. Then dense-view sinogram inpainting, image-residual learning, and image-refinement modules are performed at each iteration stage using deep neural networks. Unlike previous unrolling methods, the proposed method focuses on optimizing dense-view sinograms instead of full-view sinograms. This approach not only reduces computational resources and runtime but also minimizes the challenge for the network to perform sinogram inpainting when the sparse ratio is extremely small, thereby decreasing the propagation of estimation error from the sinogram domain to the image domain.Main results. The proposed method successfully reconstructs an image (512 × 512 pixels) from real-size (2304 × 736) projection data, with 3.39 M training parameters and an inference time of 0.09 s per slice on a GPU. The proposed method also achieves superior quantitative and qualitative results compared with state-of-the-art deep unrolling methods on datasets with sparse ratios of 1/12 and 1/18, especially in suppressing artifacts and preserving structural details. Additionally, results show that using dense-view sinogram inpainting not only accelerates the computational speed but also leads to faster network convergence and further improvements in reconstruction results.Significance. This research presents an efficient dual-domain deep unrolling technique that produces excellent results in SVCT reconstruction while requiring small computational resources. These findings have important implications for speeding up deep unrolling CT reconstruction methods and making them more practical for processing clinical CT projection data.

An AI-Based Digital Scanner for Varroa destructor Detection in Beekeeping.

Beekeeping is a crucial agricultural practice that significantly enhances environmental health and food production through effective pollination by honey bees. However, honey bees face numerous threats, including exotic parasites, large-scale transportation, and common agricultural practices that may increase the risk of parasite and pathogen transmission. A major threat is the Varroa destructor mite, which feeds on honey bee fat bodies and transmits viruses, leading to significant colony losses. Detecting the parasite and defining the intervention thresholds for effective treatment is a difficult and time-consuming task; different detection methods exist, but they are mainly based on human eye observations, resulting in low accuracy. This study introduces a digital portable scanner coupled with an AI algorithm (BeeVS) used to detect Varroa mites. The device works through image analysis of a sticky sheet previously placed under the beehive for some days, intercepting the Varroa mites that naturally fall. In this study, the scanner was tested for 17 weeks, receiving sheets from 5 beehives every week, and checking the accuracy, reliability, and speed of the method compared to conventional human visual inspection. The results highlighted the high repeatability of the measurements (R2 ≥ 0.998) and the high accuracy of the BeeVS device; when at least 10 mites per sheet were present, the device showed a cumulative percentage error below 1%, compared to approximately 20% for human visual observation. Given its repeatability and reliability, the device can be considered a valid tool for beekeepers and scientists, offering the opportunity to monitor many beehives in a short time, unlike visual counting, which is done on a sample basis.

Modeling, analysis, and optimization of random error in indirect time-of-flight camera

For indirect time-of-flight (iToF) cameras, we proposed a modeling approach focused on addressing random error. Our model characterizes random error comprehensively by detailing the propagation of error introduced by signal light, ambient light, and dark noise through phase calculation and system correction processes. This framework leverages correlations between incident light and tap responses to quantify noise impacts accurately. We then experimentally validated the theoretical model, confirming its predictive accuracy. Additionally, from a waveform design perspective, we recommend selecting an optimal duty cycle for the light waveform based on the relative intensities of ambient and signal light to effectively reduce random error.

Kinetic Mechanism and Experimental Study of Initial Temperature and Blended Gas Addition on Combustion Characteristics of Coal Bed Methane-Air

ABSTRACT Coal bed methane (CBM) is a clean energy resource, CH4 and various blends of C2H6, C2H4, CO, and H2 in different proportions were chosen to formulate a surrogate fuel for coal bed methane. The objective was to study the explosion characteristics of coal bed methane while alleviating the computational complexity associated with intricate reaction mechanisms. The explosion properties (peak explosion pressure (P max) and the maximum pressure rise rate (dp/dt) max) of the mixtures were tested. The unstretched laminar burning velocity (U l ) of the surrogate components was determined under a range of initial temperatures (T u ) (298–373 K) and blended fuel concentrations (0.4–2.0%). Two simplified kinetics models for the surrogate fuel were developed using the GRI Mech 3.0 and USC Mech II mechanisms. These models were constructed through the application of a direct relation graph with error propagation (DRGEP) and full species sensitivity analysis (FSSA). To assess the accuracy of the models, simulations based on these simplified mechanisms were compared with both the detailed mechanism and experimental data. Additionally, sensitivity analyses, utilizing the simplified mechanism, were conducted in order to identify precisely the reactions that govern the interaction dynamics between blended fuel and methane. Within the range of the experimental conditions, the results indicated that P max decreased linearly and with increasing initial temperature. (dp/dt) max was only slightly sensitive to the variation in the initial temperature. However, U l increased with increasing initial temperature, as expected. When blended fuel was added to a 7% CH4-air mixture, P max, (dp/dt)max, and unstretched U l exhibited increasing trends. An equation was derived using the experimental data, to forecast changes in U l across elevated temperatures and blended gas concentrations. The simplified mechanism model successfully replicated the results of the detailed mechanism model and demonstrated good agreement with experimental data concerning species concentrations, ignition delay times, and the U l of the coal bed methane surrogate fuel. Compared to sample 2, the chemistry of sample 1 exhibits greater sensitivity toward intermediate temperature reactions, particularly at higher temperatures.

CT material decomposition with contrast agents: Single or multiple spectral photon-counting CT scans? A simulation study.

With the widespread introduction of dual energy computed tomography (DECT), applications utilizing the spectral information to perform material decomposition became available. Among these, a popular application is to decompose contrast-enhanced CT images into virtual non-contrast (VNC) or virtual non-iodine images and into iodine maps. In 2021, photon-counting CT (PCCT) was introduced, which is another spectral CT modality. It allows for scans with more than two different detected spectra. With these systems, it becomes possible to distinguish more than two materials. It is frequently proposed to administer more than one contrast agent, perform a single PCCT scan, and then calculate the VNC images and the contrast agent maps. This may not be optimal because the patient is injected with a material, only to have it computationally extracted again immediately afterwards by spectral CT. It may be better to do an unenhanced scan followed by one or more contrast-enhanced scans. The main argument for the spectral material decomposition is patient motion, which poses a significant challenge for approaches involving two or more temporally separated scans. In this work, we assume that we can correct for patient motion and thus are free to scan the patient more than once. Our goal is then to quantify the penalty for performing a single contrast-enhanced scan rather than a clever series of unenhanced and enhanced scans. In particular, we consider the impact on patient dose and imagequality. We simulate CT scans of three differently sized phantoms containing various contrast agents. We do this for a variety of tube voltage settings, a variety of patient-specific prefilter (PSP) thicknesses and a variety of threshold settings of the photon-counting detector with up to four energy bins. The reconstructed bin images give the expectation values of soft tissue and of the contrast agents. Error propagation of projection noise into the images yields the image noise. Dose is quantified using the total CT dose index (CTDI) value of the scans. When combining multiple scans, we further consider all possible tube current (or dose) ratios between the scans. Material decomposition is done image-based in a statistical optimal way. Error propagation into the material-specific images yields the signal-to-noise ratio at unit dose (SNRD). The winning scan strategy is the one with the highest total SNRD, which is related to the SNRD of the material that has the lowest signal-to-noise ratio (SNR) among the materials to decompose into. We consider scan strategies with up to three scans and up to three materials (water W, contrast agent X and contrast agent Y). In all cases, those scan strategies yield the best performance that combine differently enhanced scans, for example, W+WX, W+WXY, WX+WXY, W+WX+WY, with W denoting an unenhanced scan and WX, WY and WXY denoting X-, Y-, and X-Y-enhanced scans, respectively. The dose efficiency of scans with a single enhancement scheme, such as WX or WXY, is far lower. The dose penalty to pay for these single enhancement strategies is about two or greater. Our findings also apply to scans with a single energy bin and thus also to CT systems with conventional, energy-integrating detectors, that is, conventional DECT. Dual source CT (DSCT) scans are preferable over single source CT scans, also because one can use a PSP on the high Kilovolt spectrum to better separate the detected spectra. For the strategies and tasks considered here, it does not make sense to simultaneously scan with two different types of contrast agents. Iodine outperforms other high Z elements in nearly allcases. Given the significant dose penalty when performing only one contrast-enhanced scan rather than a series of unenhanced and enhanced scans, one should consider avoiding the single-scan strategies. This requires to invest in the development of accurate registration algorithms that can compensate for patient and contrast agent motion between separatescans.

Equation for Calculation of Critical Current Density Using the Bean's Model with Self-Consistent Magnetic Units to Prevent Unit Conversion Errors.

This study analyzes the calculation of the critical current density Jc,mag by means of Bean's critical state model, using the equation formulated by Gyorgy et al. and other similar equations derived from it reported in the literature. While estimations of Jc,mag using Bean's model are widely performed, improper use of different equations with different magnetic units and pre-factors leads to confusion and to significant errors in the reported values of Jc,mag. In this work, a SINGLE general equation is proposed for the calculation of Jc,mag for a rectangular parallelepiped sample in perpendicular field using Bean's critical state model, underlying how the simple conversion of magnetic units can lead to a Jc,mag in the desired units, without the need to introduce any other correction or use other specific equations depending on the units of Jc,mag. In this equation, the numerical pre-factor is dimensionless, independent of the unit system used. A comparison between the expression reported in the literature is done, showing how they can lead to different results depending on the used units, and that these results can be at least one order of magnitude different from the correct results obtained with the general equation proposed in this work. This resolves all ambiguities and aligns with the correct dimensional analysis, eliminates discrepancies in the calculated Jc,mag, and will avoid further propagation of errors in the literature.

Semi-supervised learning based on distance measurement and augmentations comparison for SAR image recognition

ABSTRACT In recent years, deep learning has significantly improved the performance of Synthetic Aperture Radar Automatic Target Recognition (SAR-ATR), but the lack of large-scale labelled training data limits the further development of deep neural networks in this field. Traditionally, annotating SAR data requires manual labour by experts with specialized domain knowledge, which is costly and makes it difficult to obtain a large amount of labelled data. Semi-supervised learning (SSL) algorithms can effectively leverage unlabelled data for training, improving model performance. However, improper use of pseudo-labels can lead to error accumulation and result in performance degradation. This paper proposes an SSL algorithm for SAR image recognition that employs the following strategies to mitigate error accumulation and enhance performance: First, by combining pseudo-labelling and consistency-regularization methods, unlabelled data with different augmentations are compared in terms of classification predictions and feature vectors, improving the model representation capacity. Second, a weighted distance metric combining labelled and unlabelled data is designed to estimate model loss, guiding the model towards intra-class aggregation and inter-class dispersion in terms of data distribution. Experiments conducted on the MSTAR and OpenSARShip datasets demonstrate that the algorithm achieves superior performance in limited labelled sample scenarios, with an accuracy of over 99% when only 30 labelled data points (approximately 10%) per class are available, and an accuracy of over 85% with extremely scarce samples (only 5 labelled samples per class) on the MSTAR dataset.

Real-time, accurate, and open source upper-limb musculoskeletal analysis using a single RGBD camera — An exploratory hand-cycling study

Biomechanical biofeedback may enhance rehabilitation and provide clinicians with more objective task evaluation. These feedbacks often rely on expensive motion capture systems (∼$100000), which restricts their widespread use, leading to the development of computer vision-based methods. These methods are subject to large joint angle errors, considering the upper limb, and exclude the scapula and clavicle motion in the analysis. Our open-source approach offers a user-friendly solution for high-fidelity upper-limb kinematics using a single consumer-grade depth-sensing camera (∼$500) and includes semi-automatic skin marker labeling. Real-time biomechanical analysis, ranging from kinematics to muscle force estimation, was conducted on eight participants performing a hand-cycling motion to demonstrate the applicability of our approach on the upper limb. Markers were recorded by the depth-sensing camera and an optoelectronic camera system, considered as a reference. Muscle activity and external load were recorded using eight electromyography sensors and instrumented hand pedals, respectively. Bland-Altman analysis revealed significant agreements in the 3D markers’ positions between the two motion capture methods, with errors averaging 3.3 ± 3.9 mm. The error propagation was low for the biomechanical analysis, with joint angle differences, for example, below 5° when comparing both systems. Biofeedback from the depth-sensing camera was provided at 68 Hz. Our study introduces a novel method for using a depth-sensing camera as a low-cost motion capture solution. Results from healthy participants suggest its potential for accurate kinematic reconstruction and comprehensive upper-limb biomechanical studies. Further investigation is needed to explore its clinical applications in pathological populations.

Uncertainty Estimation Based on Error Propagation Law for Multi-Robot Pose Graph Merging

Uncertainty Estimation Based on Error Propagation Law for Multi-Robot Pose Graph Merging

Measurement Errors Propagation Laws and Schemes Comparative Study in Sensor Location Problem Based on Turning Ratios

Measurement Errors Propagation Laws and Schemes Comparative Study in Sensor Location Problem Based on Turning Ratios

EventHDR: From Event to High-Speed HDR Videos and Beyond.

Event cameras are innovative neuromorphic sensors that asynchronously capture the scene dynamics. Due to the event-triggering mechanism, such cameras record event streams with much shorter response latency and higher intensity sensitivity compared to conventional cameras. On the basis of these features, previous works have attempted to reconstruct high dynamic range (HDR) videos from events, but have either suffered from unrealistic artifacts or failed to provide sufficiently high frame rates. In this paper, we present a recurrent convolutional neural network that reconstruct high-speed HDR videos from event sequences, with a key frame guidance to prevent potential error accumulation caused by the sparse event data. Additionally, to address the problem of severely limited real dataset, we develop a new optical system to collect a real-world dataset with paired high-speed HDR videos and event streams, facilitating future research in this field. Our dataset provides the first real paired dataset for event-to-HDR reconstruction, avoiding potential inaccuracies from simulation strategies. Experimental results demonstrate that our method can generate high-quality, high-speed HDR videos. We further explore the potential of our work in cross-camera reconstruction and downstream computer vision tasks, including object detection, panoramic segmentation, optical flow estimation, and monocular depth estimation under HDR scenarios.

A modified method for measuring geometric errors of rotary axes based on sensitivity analysis and error simulation

Complicated measurement for machine tools geometric errors limits the high-accuracy machining processes. This research is aimed at measuring and identifying geometric errors of rotary axes efficiently by designing and optimizing measurement method using double ballbar. First, the conical and tangential measurement patterns are designed to reduce the installation times of the double ballbar. Secondly, the effect of installation errors is reduced by analyzing the coefficient matrix to avoid error accumulation. Thirdly, measurement trajectories affected by position-independent geometric errors are simulated and eliminated. Finally, the error measurement and validation experiments were performed on a dual-turntable five-axis machine tool, and the machining precision was improved by 34.09%-48.89%. The results indicated that the scheme can effectively identify the geometric errors of rotary axes and has the advantages of high efficiency and accuracy.

Predicting transient performance of a heavy-duty gaseous-fuelled engine using combined phenomenological and machine learning models

Decarbonizing long-haul goods transportation poses a substantial challenge. High-efficiency natural gas (NG) engines, which retain the efficiency of a diesel engine but reduce the carbon content of the fuel, offer substantial potential for near-term greenhouse gas (GHG) reductions. A fast-running model that can predict engine performance, GHG and air pollutant emissions is critical to assessing this approach for different applications and vehicle drivetrain configurations. This paper presents the development, validation and application of an engine system model that adapts GT-SUITE™’s phenomenological DI-Pulse predictive model to predict the performance and emissions of a 6-cylinder NG engine using a high pressure direct-injection combustion process. The model includes the engine air exchange system, enabling the prediction of the engine and in-cylinder conditions and overall performance over transient drive cycles. The engine model with a fixed set of calibration parameters captures the complex high-pressure direct injection combustion process and generates time-resolved parameters that are fed into a coupled machine learning model to predict emissions, including nitrogen oxide (NOx) and methane (CH4) emissions. While the 1-D model’s predictions for CH4 were not accurate, coupling the 1-D engine model with a machine learning model has been shown to substantially improve the estimation of CH4 emissions and allow accurate prediction of engine total GHG emissions over different duty cycles. The model has been validated using transient engine dynamometer data and is then applied to assess performance and emissions over several regulatory and real-world long-haul drive cycles. The model showed an average error of less than 5% in steady operation. Cumulative errors of NOx and CH4 emissions in studied cycles were also less than 10%. The results showed that CH4 share in total GHG emissions ranges from 0.2% to 1.4% over various drive cycles. By predicting engine performance and emissions, the developed combined model has considerable potential for use in engine evaluation studies, especially when combined with new technologies across different duty cycles.

Research on Spaceborne Neural Network Accelerator and Its Fault Tolerance Design

To meet the high-reliability requirements of real-time on-orbit tasks, this paper proposes a fault-tolerant reinforcement design method for spaceborne intelligent processing algorithms based on convolutional neural networks (CNNs). This method is built on a CNN accelerator using Field-Programmable Gate Array (FPGA) technology, analyzing the impact of Single-Event Upsets (SEUs) on neural network computation. The accelerator design integrates data validation, Triple Modular Redundancy (TMR), and other techniques, optimizing a partial fault-tolerant architecture based on SEU sensitivity. This fault-tolerant architecture analyzes the hardware accelerator, parameter storage, and actual computation, employing data validation to reinforce model parameters and spatial and temporal TMR to reinforce accelerator computations. Using the ResNet18 model, fault tolerance performance tests were conducted by simulating SEUs. Compared to the prototype network, this fault-tolerant design method increases tolerance to SEU error accumulation by five times while increasing resource consumption by less than 15%, making it more suitable for spaceborne on-orbit applications than traditional fault-tolerant design approaches.

Optimizing Automated Hematoma Expansion Classification from Baseline and Follow-Up Head Computed Tomography

Hematoma expansion (HE) is an independent predictor of poor outcomes and a modifiable treatment target in intracerebral hemorrhage (ICH). Evaluating HE in large datasets requires segmentation of hematomas on admission and follow-up CT scans, a process that is time-consuming and labor-intensive in large-scale studies. Automated segmentation of hematomas can expedite this process; however, cumulative errors from segmentation on admission and follow-up scans can hamper accurate HE classification. In this study, we combined a tandem deep-learning classification model with automated segmentation to generate probability measures for false HE classifications. With this strategy, we can limit expert review of automated hematoma segmentations to a subset of the dataset, tailored to the research team’s preferred sensitivity or specificity thresholds and their tolerance for false-positive versus false-negative results. We utilized three separate multicentric cohorts for cross-validation/training, internal testing, and external validation (n = 2261) to develop and test a pipeline for automated hematoma segmentation and to generate ground truth binary HE annotations (≥3, ≥6, ≥9, and ≥12.5 mL). Applying a 95% sensitivity threshold for HE classification showed a practical and efficient strategy for HE annotation in large ICH datasets. This threshold excluded 47–88% of test-negative predictions from expert review of automated segmentations for different HE definitions, with less than 2% false-negative misclassification in both internal and external validation cohorts. Our pipeline offers a time-efficient and optimizable method for generating ground truth HE classifications in large ICH datasets, reducing the burden of expert review of automated hematoma segmentations while minimizing misclassification rate.

DFE Error Propagation Mitigation Using Feed Forward Equalizer and Forward Error Correction in 25 Gb/s NRZ Transceiver

In order to mitigate DFE error propagation to reduce bit error rate (BER), a 25[Formula: see text]Gb/s Non-Return-To-Zero (NRZ) transceiver including a five-tap decision feedback equalizer (DFE) combined with a dedicated feed forward equalizer (FFE) was first constructed and was found to be helpful in improving signal integrity and reduce errors in high-speed data transmission, which can lead to more reliable and efficient communication. In this transceiver, the effect of multi-tap FFE on error probability with different burst error run lengths was evaluated through a cumulative-probability distribution model. Then, forward error correction (FEC) with different correction capabilities can be also employed to correct burst errors for a better BER performance. Probability-based BER with single/double burst error correcting FEC was derived and a joint optimization of FFE and FEC can be further explored to achieve the target BER. Results show that compared with two-tap FFE, FFE with a post-cursor tap can obtain a BER improvement of 1.5[Formula: see text]dB at the BER of [Formula: see text], and also enlarge the horizontal opening degree of DFE, which is 0.74 unit interval (UI). Double burst error correcting FEC obtaining 2.2[Formula: see text]dB coding gain at the BER of [Formula: see text] outperforms single one. When the FEC error correcting capability is 16, the combination of double burst error correcting FEC and three-tap FFE can reduce BER from [Formula: see text] to approximately [Formula: see text]. FEC with a larger error correction capability is more likely to achieve the target BER and reduce the complexity of the combination equalizer while ensuring a better equalization performance. The proposed NRZ transceiver, as a part of the physical interface, can be applied to high-speed interconnection scenarios, such as Ethernet, PCIe, and Chiplet.