Loss Of Accuracy Research Articles

A robust iterative algorithm for SINS/USBL integrated navigation based on dual hydrophone differential model

Aiming at the complex underwater environment which leads to large measurement errors and outliers in ultrashort baseline, a robust iterative Kalman filtering algorithm based on the differential model of dual hydrophones is proposed in this paper. Firstly, to achieve accurate navigation under large initial misalignment angles, the algorithm builds an inertial navigation error equation based on the right invariant error definition of Lie groups. Then a dual hydrophone differential model is proposed based on SINS/USBL tightly coupled. It successfully improves the accuracy of integrated navigation and significantly mitigating the issue of SINS/USBL positioning accuracy degradation caused by USBL slant distance noise that grows with distance. Finally, a robust iterative Kalman filter based on the dual hydrophone differential model is proposed to counter the threat to the vehicle’s safe operation and the accuracy loss caused by outliers. The algorithm’s superiority and effectiveness are verified through simulation and sea trial.

Finite-time output-feedback fault tolerant adaptive fuzzy control framework for a class of MIMO saturated nonlinear systems

This paper presents a finite-time fault-tolerant fuzzy adaptive controller for uncertain interconnected nonlinear systems in strict-feedback form. The controller addresses input saturation, state-dependent actuator faults, external disturbances, and unavailable states for measurement. To avoid any unexpected alteration due to failures, the proposed control scheme design addresses all types of actuator faults that are bias, drift, and loss of accuracy as additive faults, as well as a multiplicative fault that results in loss of effectiveness with nonaffine state-dependency. Restrictions on control gains and unmatched interconnections are eliminated, and the explosion complexity caused by recursive back-stepping designs is avoided. Fuzzy Systems approximate the unknown ideal control laws along side the acutator faults and disturbances. The stability analysis guarantees that tracking errors converge to a small compact set around the origin in finite time. Simulation results on a quad-rotor UAV and a mathematical system confirm the effectiveness of the proposed approach.

Multiple fairness criteria in decision tree learning

The use of algorithmic decision-making systems based on machine learning models has led to a need for fair (unbiased) and explainable classification outcomes. In particular, machine learning algorithms can encode biases, which might result in discriminatory decisions for certain groups such as gender, race, or age. Although a number of works on decision tree learning have been proposed to decrease the chance of discrimination, they usually focus on the use of a single fairness metric. In general, creating a model based on a single fairness metric is not a sufficient way to mitigate discrimination since bias can originate from various sources—e.g., the data itself or the optimization process. In this paper, we propose a novel decision tree learning process that utilizes multiple fairness metrics to address both group and individual discrimination. This is achieved by extending the attribute selection procedure to consider not only information gain but also gain in fairness. Computational experiments on fourteen different datasets with various sensitive features demonstrate that the proposed Fair-C4.5 models improve fairness without a loss in predictive accuracy when compared to the well-known C4.5 and the fairness-aware FFTree algorithms.

Adaptive feature alignment for adversarial training

Recent studies reveal that Convolutional Neural Networks (CNNs) are typically vulnerable to adversarial attacks. Many adversarial defense methods have been proposed to improve the robustness against adversarial samples. Moreover, these methods can only defend adversarial samples of a specific strength, reducing their flexibility against attacks of varying strengths. Moreover, these methods often enhance adversarial robustness at the expense of accuracy on clean samples. In this paper, we first observed that features of adversarial images change monotonically and smoothly w.r.t the rising of attacking strength. This intriguing observation suggests that features of adversarial images with various attacking strengths can be approximated by interpolating between the features of adversarial images with the strongest and weakest attacking strengths. Due to the monotonicity property, the interpolation weight can be easily learned by a neural network. Based on the observation, we proposed the adaptive feature alignment (AFA) that automatically align features to defense adversarial attacks of various attacking strengths. During training, our method learns the statistical information of adversarial samples with various attacking strengths using a dual batchnorm architecture. In this architecture, each batchnorm process handles samples of a specific attacking strength. During inference, our method automatically adjusts to varying attacking strengths by linearly interpolating the dual-BN features. Unlike previous methods that need to either retrain the model or manually tune hyper-parameters for a new attacking strength, our method can deal with arbitrary attacking strengths with a single model without introducing any hyper-parameter. Additionally, our method improves the model robustness against adversarial samples without incurring much loss of accuracy on clean images. Experiments on CIFAR-10, SVHN and tiny-ImageNet datasets demonstrate that our method outperforms the state-of-the-art under various attacking strengths and even improve accuracy on clean samples. Code will be made open available upon acceptance.

Robust Fuzzy Adaptive Fault‐Tolerant Control for a Class of Second‐Order Nonlinear Systems

ABSTRACTIn this brief, an optimal active fuzzy fault‐tolerant control (OAFFTC) scheme is proposed for a class of unknown perturbed multi‐input multi‐output (MIMO) nonlinear systems with nonaffine nonlinear actuator faults and time‐varying sensor faults. The control strategy can handle automatically (online updating) with three additives (bias, drift, and loss of accuracy) along with one multiplicative (loss of effectiveness) sensor faults and nonlinear state‐dependent actuator faults. In order to deal with uncertain system dynamics, sensor and actuator faults, and external disturbances, fuzzy systems (FSs) and backstepping techniques were combined to provide the adaptive control term as well as a robust control term. Butterworth filter is used to get rid the algebraic loop problem. The suggested robust term, can deal with approximation errors of the fuzzy systems (FSs). To automatically optimize the adaptive parameters and the starting conditions, particle swarm optimization (PSO) approach is introduced. The Lyapunov approach is employed to demonstrate the closed‐loop system's stability. To assess the efficacy and correctness of the suggested scheme, quadrotor dynamic model is performed on the simulation part.

Neighbor patches merging reduces spatial redundancy to accelerate vision transformer

Vision Transformers (ViTs) deliver outstanding performance but often require substantial computational resources. Various token pruning methods have been developed to enhance throughput by removing redundant tokens; however, these methods do not address the peak memory consumption, which remains equivalent to that of the unpruned networks. In this study, we introduce Neighbor Patches Merging (NEPAM), a method that significantly reduces the maximum memory footprint of ViTs while pruning tokens. NEPAM targets spatial redundancy within images and prunes redundant patches at the onset of the model, thereby achieving the optimal throughput-accuracy trade-off without fine-tuning. Experimental results demonstrate that NEPAM can accelerate the inference speed of the Vit-Base-Patch16-384 model by 25% with a negligible accuracy loss of 0.07% and a notable 18% reduction in memory usage. When applied to VideoMAE, NEPAM doubles the throughput with a 0.29% accuracy loss and a 48% reduction in memory usage. These findings underscore the efficacy of NEPAM in mitigating computational requirements while maintaining model performance.

Advanced Dual-Input Artificial Optical Synapse for Recognition and Generative Neural Network

Perovskite materials have emerged as leading candidates for optical synaptic devices due to their superior photosensitivity and tunable optoelectronic properties. However, the practical application of perovskite-based optoelectronic synaptic transistors has been hindered by issues of poor stability and high toxicity. This study developed an artificial synaptic thin film transistor (TFT) based on a Cs2AgBiBr6/InOx bilayer. The device demonstrated synaptic behavior under optoelectronic hybrid stimulation (largest wavelength ~520nm), such as excitatory postsynaptic current (EPSC), inhibitory postsynaptic current (IPSC), paired-pulse facilitation (PPF), spike-timing-dependent plasticity (STDP), short-term memory (STM) and long-term memory (LTM). Notably, the “light writing and voltage erasing” characteristics of these devices could be utilized to construct a convolutional neural network (CNN) classifier for the CIFAR-10 dataset, demonstrating noise tolerance close to the human eye. The loss in recognition accuracy was within 1% when Gaussian white noise and salt pepper noise were added. Furthermore, these devices exhibited great potential in Generative Adversarial Networks (CycleGAN), with the generated image quality achieving levels of 0.637 and 0.715 in Improved Perceptual Image Processing System (IPIPS) and Structural Similarity Index (SSIM) evaluation metrics for the zebra dataset, indicating good image quality. This study indicates that our Cs2AgBiBr6-based artificial optical synaptic thin-film transistors (TFTs) are promising for sensing and in-memory computing applications.

COVERater-A Free Application for Training Researchers to Accurately Estimate Species Cover in Terrestrial and Aquatic Ecosystems.

Visual estimates of cover are widely used among ecologists, from describing vegetation communities to tracking and monitoring species' abundance. However, despite the known bias associated with visual estimates, no standardised training is available to improve these measurements. We developed a free online training tool, the COVERater, that effectively teaches users to visually estimate the percent cover of species in a variety of ecosystems (including alpine heath, arid lands, coral reefs, temperate reefs and wetlands). Prior to training, users with prior professional experience estimated species cover to an average inaccuracy of 5.2%, while users with no experience estimated cover to an average inaccuracy of 7.6%. COVERater training took an average of 31 min and 68 images, and reduced the estimate inaccuracy of users with no prior experience to 5.2%. There was no significant loss of estimate accuracy over 100 days following training. The COVERater can be used anywhere in the world, by data collectors of all experience levels, for projects spanning all spatial scales. By providing researchers with standardised training, our application can reduce variation in cover estimates that arise from human biases, allowing for comparable estimates across global collaborative projects and data syntheses. We encourage all relevant scientists to include COVERater training in their protocols to quantify cover with greater accuracy, improve the veracity of their results and make better inferences about our biosphere.

Ultrashort Oncologic Whole-Body [18F]FDG Patlak Imaging Using LAFOV PET.

Methods to shorten [18F]FDG Patlak PET imaging procedures ranging from 65-90 to 20-30 min after injection, using a population-averaged input function (PIF) scaled to patient-specific image-derived input function (IDIF) values, were recently evaluated. The aim of the present study was to explore the feasibility of ultrashort 10-min [18F]FDG Patlak imaging at 55-65 min after injection using a PIF combined with direct Patlak reconstructions to provide reliable quantitative accuracy of lung tumor uptake, compared with a full-duration 65-min acquisition using an IDIF. Methods: Patients underwent a 65-min dynamic PET acquisition on a long-axial-field-of-view (LAFOV) Biograph Vision Quadra PET/CT scanner. Subsequently, direct Patlak reconstructions and image-based (with reconstructed dynamic images) Patlak analyses were performed using both the IDIF (time to relative kinetic equilibrium between blood and tissue concentration (t*) = 30 min) and a scaled PIF at 30-60 min after injection. Next, direct Patlak reconstructions were performed on the system console using only the last 10 min of the acquisition, that is, from 55 to 65 min after injection, and a scaled PIF using maximum crystal ring difference settings of both 85 and 322. Tumor lesion and healthy-tissue uptake was quantified and compared between the differently obtained parametric images to assess quantitative accuracy. Results: Good agreement was obtained between direct- and image-based Patlak analyses using the IDIF (t* = 30 min) and scaled PIF at 30-60 min after injection, performed using the different approaches, with no more than 8.8% deviation in tumor influx rate value (Ki ) (mean difference ranging from -0.0022 to 0.0018 mL/[min × g]). When direct Patlak reconstruction was performed on the system console, excellent agreement was found between the use of a scaled PIF at 30-60 min after injection versus 55-65 min after injection, with 2.4% deviation in tumor Ki (median difference, -0.0018 mL/[min × g]; range, -0.0047 to 0.0036 mL/[min × g]). For different maximum crystal ring difference settings using the scan time interval of 55-65 min after injection, only a 0.5% difference (median difference, 0.0000 mL/[min × g]; range, -0.0004 to 0.0013 mL/[min × g]) in tumor Ki was found. Conclusion: Ultrashort whole-body [18F]FDG Patlak imaging is feasible on an LAFOV Biograph Vision Quadra PET/CT system without loss of quantitative accuracy to assess lung tumor uptake compared with a full-duration 65-min acquisition. The ultrashort 10-min direct Patlak reconstruction with PIF allows for its implementation in clinical practice.

A Novel Approach Integrating the Method of Characteristics with Large Time-Step Scheme (MOC-LTS) for Efficient Transient Flow Simulation in Liquid Pipelines

Water hammer incidents pose a significant risk to the safety and stability of pipeline operations. Therefore, rapid and precise transient flow simulation is essential for efficiently developing scientific water hammer control strategies. Nevertheless, the prevailing transient flow simulation methods for liquid pipelines predominantly employ explicit schemes to solve transient flow control equations, necessitating adherence to the stability criterion of Courant-Friedrichs-Lewy (CFL) ≤ 1. This results in limited time step sizes, which in turn constrains computational efficiency, particularly when simulating hydraulic behavior in large-scale pipeline networks. In this paper, a novel approach integrating the method of characteristics (MOC) with a large time-step scheme (LTS) is proposed to enable rapid and accurate simulation of transient flow in liquid pipelines. The proposed approach discretizes the computational domain into contiguous control volumes, ategorizing them as either boundary or internal control volumes depending on whether they are affected by boundary conditions within a large time step. For internal control volumes, an LTS scheme based on the first-order Godunov format is employed to improve computational efficiency. For boundary control volumes, MOC is applied iteratively to accurately capture boundary dynamics until the simulated time matches that of the internal control volumes. Quantitative analysis is conducted to verify the performance of the proposed approach. The results confirm that the proposed approach outperforms the MOC in computational efficiency while maintaining minimal accuracy loss.

Information hiding using approximate POSIT representation

AbstractPOSIT is a numerical system consists of sign, regime, exponential, and fraction bits to overcome some limitations of the IEEE‐754 floating point (FP) representation. This work proposes a technique to hide critical information in the least significant bits (LSBs) of the POSIT FP representation by exploiting approximate computing (AC). The proposed technique, called information hiding using POSIT and LSB (IHUPL), explores the opportunities offered by both the POSIT representation and the AC to hide information with a minimum loss in accuracy and other performance metrics. IHUPL offers two options for hiding information: either by embedding one digit of each character into each pixel or by embedding all the digits of the character at the same time into each pixel of the alpha‐red, green, blue, or black/white image. Experimental results are evaluated for benchmark images and showed that IHUPL enhanced the accuracy of embedding data into LSB of POSIT FP by an average of 5%, the image quality improvement rates of IHUPL are 19% and 16% for options 1 and 2, respectively. Besides the encoding method using IHUPL, the paper outlines an extraction decoding technique that saves the original replacement bits in a key‐image to recover the hidden security message.

Modeling viscosity of commercial glass melts: Adam–Gibbs equation and Gaussian process regression

ABSTRACT A statistically designed property database of six commercial glass families is leveraged to develop models relating viscosity to temperature and composition. Specialized models for each glass family were designed based on the Adam–Gibbs equation. We show that the whole range of viscosity data of one glass family from 100 to 1012 Pa s can be accurately modeled using one composition-dependent and two composition-independent parameters when glass transition temperature and viscosity at that temperature are independently determined. Generalization to a broader composition space using the same approach leads to a loss of accuracy compared to specialized models. However, this can be mostly negated when Gaussian process regression is used to estimate the compositional dependence of configuration entropy and fragility exponent parameters of the Adam–Gibbs equation.

An experimental design-based approach for modelling of weapon engagement zone of an air-to-air missile

ABSTRACT It is vital for pilots to have precise information about missile ranges during air-to-air combat. The Weapon Engagement Zone (WEZ), or Dynamic Launch Zone (DLZ) for air missions, represents these ranges based on the flight conditions of both the launching and target aircraft. Generating accurate, real-time WEZ functions requires high-fidelity simulations under various engagement scenarios. Classical approaches often require numerous simulations due to uncertainty in sample requirements and complexities like nonlinearities. To address this, the study proposes a method that treats WEZ modeling as a computer experiment, using a surrogate model developed through sequential experimental design and deep neural networks (DNN). This iterative approach optimizes the number of simulations needed, minimizing the loss of model accuracy. The method’s effectiveness is validated by comparing it to classical factorial designs, showing that the proposed approach achieves similar accuracy with significantly fewer sample points, making it a more efficient solution for WEZ modeling.

Fatigue delamination damage analysis in composite materials through a rule of mixtures approach

The present study investigates delamination damage initiation and propagation within a homogenization theory of mixtures, using the concept of virtual layers and virtual interfaces. It eliminates spatial discretization of layers, introducing a resultant damage variable to capture structure’s bulk response under both monotonic and cyclic loads. Fatigue-induced deterioration is classified into sub-critical, critical, and over-critical stages based on interfacial stresses. Calibration is conducted employing the widely-available Wöhler curves for each loading mode independently. An advance-in-time strategy is included in the model to enhance the simulation speed. The reliability of the approach is assessed for crack initiation and propagation separately through standard test coupons, showing good correlation with experimental data in mode I, mode II, and mixed-mode loading conditions. Depending on the calibration procedure adopted, the model is applicable to a wide range of stress ratios. In addition, it could be integrated into any standard finite element framework using the desired number of elements through the thickness regardless of the physical amount of layers. This allows easy modification of stacking sequences or the number of layers within the constitutive law without mesh structure changes, facilitating simulation of large-scale composite laminates with minimal accuracy loss and reduced computational costs.

Local differential privacy federated learning based on heterogeneous data multi-privacy mechanism

Federated learning enables the development of robust models without accessing users data directly. However, recent studies indicate that federated learning remains vulnerable to privacy leakage. To address this issue, local differential privacy mechanisms have been incorporated into federated learning. Nevertheless, local differential privacy will reduce the availability of data. To explore the balance between privacy budgets and data availability in federated learning, we propose federated learning for clustering hierarchical aggregation with adaptive piecewise mechanisms under multiple privacy-FedAPCA as a way to balance the relationship between privacy preservation and model accuracy. First, we introduce an adaptive piecewise mechanism that dynamically adjusts perturbation intervals based on the data ranges across different layers of the model, ensuring minimized perturbation variance while maintaining the same level of privacy. Second, we propose two dynamic privacy budget allocation methods, which are allocating the privacy budget based on global accuracy and global loss, and allocating the privacy budget based on local accuracy and loss, to ensure that better model accuracy can be achieved under the same privacy budget. Finally, we propose a clustering hierarchical aggregation method in the model aggregation stage, and the model is updated and aggregated after the unbiased estimation of the disturbance in each cluster according to the variance of each layer. FedAPCA improves the balance between privacy preservation and model accuracy. Our experimental results, comparing FedAPCA with the SOTA multi-privacy local differential privacy federated learning frameworks on the MNIST and CIFAR-10 datasets, demonstrate that FedAPCA improves model accuracy by 1%–2%.

RCYOLO: an innovative surface defect detection model for magnetic blocks integrating RevCol and Yolov8s

ABSTRACT The surface defect detection of magnetic blocks faces challenges such as low accuracy, high missed detection rates, and weak anti-interference capabilities. Additionally, existing models are often large in size and require significant computational resources. To address these issues, this article integrated the Yolov8 framework and designed the RCYOLO model to improve detection performance and resource utilization. Firstly, we used RevCol as the backbone, and explored its optimal number of sub-networks to reduce model parameters and complexity, as well as improve model robustness. Secondly, we removed some deep structures from Yolov8s to simplify the model, making it easier to deploy. Afterwards, we developed a more efficient feature extraction module, C2f_SA, and replaced all existing modules with it to compensate for accuracy loss caused by simplification and improve overall detection performance. Finally, we optimized reg_max to explore the best minimum value for detecting defects in magnetic blocks to reduce unnecessary computational burden. Experimental results show that compared to Yolov8s, although RCYOLO’s mAP0.5 was reduced by 0.2%, it significantly reduced the model parameters, FLOPs, and model size by 51.64%, 30.31%, and 50.22%, respectively, achieving a better balance between accuracy and resources. Moreover, RCYOLO demonstrated good detection performance compared to other improved models.

Computer-aided diagnosis of pituitary microadenoma on dynamic contrast-enhanced MRI based on spatio-temporal features

Computer-aided diagnosis (CAD) of pituitary microadenoma (PM) can assist doctors in decision-making, leading to improved lesion detection rates and diagnostic accuracy. However, the performance of existing CAD methods for PM detection has been hindered by the difficulty in obtaining high-quality segmentation results. This is primarily due to the small size of PM lesions and the relatively low resolution of Magnetic Resonance Imaging (MRI) images. To address these challenges, this paper proposes a new medical image detection and segmentation model based on spatio-temporal information. The proposed model aims to addresses the disease classification of PM by designing a network module based on multi-scale feature fusion. This module ensures comprehensive extraction of target semantic information while retaining clear spatial information, achieving classification from dynamic contrast-enhanced MRI(DCE-MRI) to identify positive PM samples. For the lesion segmentation of PM, after ROI Align alignment, the model further adds a semantic segmentation module named Dual-path Semantic Segmentation Module (DSSM) behind the mask head and classification head. This module captures more precise spatio-temporal semantic information, reducing accuracy loss and achieving pituitary segmentation. Finally, leveraging the results of pituitary detection, a feature pyramid network (FPN) layer is redesigned named Reuse Underlying Information Module (RUIM) to reuse low-level information, enhancing the detection capability for PM and thus achieving precise object detection and segmentation. The proposed model achieves an accuracy of 97.10% for PM, mAP of 50.24%, which is superior to multiple representative deep models for medical data. The code is available at https://github.com/BUCT-IUSRC/Research__PM-CAD.

EdgeSyn: Privacy-preserving Data Publishing on Edge Network over Infinite Multimedia Data Stream

To privately publish sensitive multimedia data in an edge network with fog devices, one of the best privacy-preserving solutions is to use differential privacy (DP) mechanisms. However, existing DP data publication mechanisms for the infinite data stream of edge networks mainly focus on publishing data with specific types of data or a set of predetermined queries. This approach is not suitable for multimedia data with numerous features that require a more flexible data publishing mechanism. In this paper, we propose EdgeSyn, a novel mechanism for accurately publishing multimedia data over infinite data streams in an edge network. It allocates privacy budgets with a sliding window, adopting data synthesis mechanisms to support dynamic publishing without loss of accuracy. In more detail, EdgeSyn addresses the limitations associated with data types in prior data stream publishing approaches and introduces a privacy budget management strategy that optimally allocates budgets for the implementation of data synthesis mechanisms over an infinite data stream. The experimental results show that EdgeSyn performs well under different privacy budgets and various lengths of active windows.

The ATOM-Seq sequence capture panel can accurately predict microsatellite instability status in formalin-fixed tumour samples, alongside routine gene mutation testing

Microsatellite instability (MSI) occurs across a number of cancers and is associated with different clinical characteristics when compared to microsatellite stable (MSS) cancers. As MSI cancers have different characteristics, routine MSI testing is now recommended for a number of cancer types including colorectal cancer (CRC). Using gene panels for sequencing of known cancer mutations is routinely performed to guide treatment decisions. By adding a number of MSI regions to a small gene panel, the efficacy of simultaneous MSI detection in a series of CRCs was tested. Tumour DNA from formalin-fixed, paraffin-embedded (FFPE) tumours was sequenced using a 23-gene panel kit (ATOM-Seq) provided by GeneFirst. The mismatch repair (MMR) status was obtained for each patient from their routine pathology reports, and compared to MSI predictions from the sequencing data. By testing 29 microsatellite regions in 335 samples the MSI status was correctly classified in 314/319 samples (98.4% concordance), with sixteen failures. By reducing the number of regions in silico, comparable performance could be reached with as few as eight MSI marker positions. This test represents a quick, and accurate means of determining MSI status in FFPE CRC samples, as part of a routine gene mutation assay, and can easily be incorporated into a research or diagnostic setting. This could replace separate mutation and MSI tests with no loss of accuracy, thus improving testing efficiency.

Lazy Resampling: Fast and information preserving preprocessing for deep learning

Background and Objective:Preprocessing of data is a vital step for almost all deep learning workflows. In computer vision, manipulation of data intensity and spatial properties can improve network stability and can provide an important source of generalisation for deep neural networks. Models are frequently trained with preprocessing pipelines composed of many stages, but these pipelines come with a drawback; each stage that resamples the data costs time, degrades image quality, and adds bias to the output. Long pipelines can also be complex to design, especially in medical imaging, where cropping data early can cause significant artifacts. Methods:We present Lazy Resampling, a software that rephrases spatial preprocessing operations as a graphics pipeline. Rather than each transform individually modifying the data, the transforms generate transform descriptions that are composited together into a single resample operation wherever possible. This reduces pipeline execution time and, most importantly, limits signal degradation. It enables simpler pipeline design as crops and other operations become non-destructive. Lazy Resampling is designed in such a way that it provides the maximum benefit to users without requiring them to understand the underlying concepts or change the way that they build pipelines. Results:We evaluate Lazy Resampling by comparing traditional pipelines and the corresponding lazy resampling pipeline for the following tasks on Medical Segmentation Decathlon datasets. We demonstrate lower information loss in lazy pipelines vs. traditional pipelines. We demonstrate that Lazy Resampling can avoid catastrophic loss of semantic segmentation label accuracy occurring in traditional pipelines when passing labels through a pipeline and then back through the inverted pipeline. Finally, we demonstrate statistically significant improvements when training UNets for semantic segmentation. Conclusion:Lazy Resampling reduces the loss of information that occurs when running processing pipelines that traditionally have multiple resampling steps and enables researchers to build simpler pipelines by making operations such as rotation and cropping effectively non-destructive. It makes it possible to invert labels back through a pipeline without catastrophic loss of accuracy.A reference implementation for Lazy Resampling can be found at https://github.com/KCL-BMEIS/LazyResampling. Lazy Resampling is being implemented as a core feature in MONAI, an open source python-based deep learning library for medical imaging, with a roadmap for a full integration.