Image Error Research Articles

Design Method of Freeform Surface Optical Systems with Low Coupling Position Error Sensitivity.

Freeform off-axis reflective systems are significantly more difficult to align and assemble owing to their asymmetric surface shapes and system structures. In this study, a freeform surface system design method with low coupling position error sensitivity (FCPESM) was proposed. First, we established a mathematical model of a reflective system when it was perturbed by coupling position errors and used the clustering-microelement method to establish the coupling error sensitivity evaluation function. The evaluation function was then applied to the design process of a freeform surface off-axis three-mirror optical system. The results showed that the FCPESM optical design method can significantly relax the assembly tolerance requirements of optical systems on the basis of ensuring image performance. In this study, the reflective system was perturbed by tilt and decenter simultaneously, and the disturbance mechanism of position errors on optical systems was further improved. Through this research, freeform surface systems with both image performance and error sensitivity can be obtained, which makes freeform off-axis reflective systems with better engineering realizability.

P-203 Training of an artificial intelligence (AI) model to automatically detect nuclear errors during first cleavages

Abstract Study question Can an artificial intelligence (AI) algorithm be developed to automatically detect nuclear errors through dynamic images of first embryonic cleavages? Summary answer The AI model showed an accuracy and precision of above 60% in automatically detecting multinucleation in dynamic images. What is known already Implementation of dynamic monitoring by time-lapse (TL) incubators intends to enhance embryo selection, potentially leveraging algorithmic assistance. During embryonic development, certain morphological nuances, often overlooked in intermittent evaluations, become more apparent through TL monitoring. Some of these features are related to faulty karyokinesis in blastomeres, recognized as nuclear errors, which are linked to aneuploidy and reduced live-birth potential. Despite their significance in embryo selection, existing computer imaging AI models lack the ability to automatically detect these features. This research aims to develop the YOLO AI model’s ability to autonomously identify nuclear errors in dynamic embryonic cleavage imaging. Study design, size, duration Videos of 364 embryos recorded with four different TL incubators (Geri-Genea BiomedX, MIRI-ESCO, EmbryoScope/EmbryoScopePlus-Vitrolife) were randomly sampled from retrospective data. Videos started around the time of pronuclei fading (tPNf) and ceased at t5 (from 20hpi-50hpi). Each frame of each video was annotated by a senior embryologist. A total of 13246 frames were gathered with a resolution of 512x512 pixels, from which 9411 were used to train a YOLO model, 1977 for validation and test phases. Participants/materials, setting, methods For each frame, the embryologist outlined a box around the single nucleus observed in a blastomere and labelled it as “Normal”, and as “Multinucleated” if more than one nucleus or micronuclei were seen. 8% of the training dataset was composed of frames with no visible nucleus and 19% contained at least one multinucleated cell. The model thus further detected the presence of nuclei in a frame and classified it as “Normal” or “Multinucleated” when found. Main results and the role of chance Inferences and metrics were computed at the level of the embryo. No images of test videos were used in training. An embryo was considered “Normal” if no multinucleated cells were detected anywhere during the video. If at least one nuclear error was labelled, the embryo was considered as “Multinucleated”. To get a prediction at the embryo level, we considered the class predicted with the highest confidence during the video by the model. Another rule was to consider the embryo as “Multinucleated” if at least one cell was predicted as “Multinucleated” with a confidence above 0.5. When evaluated on the test dataset, the AI model reached 67% accuracy at detecting multinucleated embryos, with a precision of 63%, correctly identifying 61% of the multinucleated embryos. Conversely, when the model predicted that the embryo did not have any nuclear errors, it was right 61% of the time. Limitations, reasons for caution To reach the current set of results, the present model underwent previous multiple interfaces using several hundreds of embryos, all only trained by a single experienced embryologist. The model must undergo a cross-validation involving embryologists with varying levels of experience to assess its performance with a high degree of precision. Wider implications of the findings Continuous training and refinement of algorithms, informed by the accumulated knowledge of impacting features, play a crucial role in the evolution of automated systems. As these methods evolve, they will further refine embryo selection, improve outcomes, and contribute to the ongoing advancement of IVF. Trial registration number not applicable

Assessing Laterality Errors in Radiology: Comparing Generative Artificial Intelligence and Natural Language Processing

PurposeWe compared the performance of generative artificial intelligence (AI) (Augmented Transformer Assisted Radiology Intelligence [ATARI, Microsoft Nuance, Microsoft Corporation, Redmond, Washington]) and natural language processing (NLP) tools for identifying laterality errors in radiology reports and images. MethodsWe used an NLP-based (mPower, Microsoft Nuance) tool to identify radiology reports flagged for laterality errors in its Quality Assurance Dashboard. The NLP model detects and highlights laterality mismatches in radiology reports. From an initial pool of 1,124 radiology reports flagged by the NLP for laterality errors, we selected and evaluated 898 reports that encompassed radiography, CT, MRI, and ultrasound modalities to ensure comprehensive coverage. A radiologist reviewed each radiology report to assess if the flagged laterality errors were present (reporting error—true-positive) or absent (NLP error—false-positive). Next, we applied ATARI to 237 radiology reports and images with consecutive NLP true-positive (118 reports) and false-positive (119 reports) laterality errors. We estimated accuracy of NLP and generative AI tools to identify overall and modality-wise laterality errors. ResultsAmong the 898 NLP-flagged laterality errors, 64% (574 of 898) had NLP errors and 36% (324 of 898) were reporting errors. The text query ATARI feature correctly identified the absence of laterality mismatch (NLP false-positives) with a 97.4% accuracy (115 of 118 reports; 95% confidence interval [CI] = 96.5%-98.3%). Combined vision and text query resulted in 98.3% accuracy (116 of 118 reports or images; 95% CI = 97.6%-99.0%), and query alone had a 98.3% accuracy (116 of 118 images; 95% CI = 97.6%-99.0%). ConclusionThe generative AI-empowered ATARI prototype outperformed the assessed NLP tool for determining true and false laterality errors in radiology reports while enabling an image-based laterality determination. Underlying errors in ATARI text query in complex radiology reports emphasize the need for further improvement in the technology.

Encrypted image processing using compression and reversible data hiding

Reversible data hiding within encrypted images reversible data hiding in encrypted images (RDH-EI) is a highly effective technique for image processing in the field of encryption. This paper, propose a RDH-EI technique, which utilizes bit-plane compression and various image scanning directions to generate vacant space for data embedding, referred to as vacating room. Initially, the prediction error of the pre-processed image is computed. Subsequently, each bit-plane image is converted into a bit-stream by following the pixel scan order employed before compression. The compressed image is then encrypted employing a stream cipher. Through the process of substitution, the secret data and additional information are incorporated into the acquired image without any knowledge of the original content or the encrypted key. Finally, the generated image is transmitted or archived. The experiments provide evidence that the proposed method surpasses the most advanced methods currently available.

FLsM: Fuzzy Localization of Image Scenes Based on Large Models

This article primarily focuses on the study of image-based localization technology. While traditional methods have made significant advancements in technology and applications, the emerging field of visual image-based localization technology demonstrates tremendous potential for research. Deep learning has exhibited a strong performance in image processing, particularly in developing visual navigation and localization techniques using large-scale visual models. This paper introduces a sophisticated scene image localization technique based on large models in a vast spatial sample environment. The study involved training convolutional neural networks using millions of geographically labeled images, extracting image position information using large model algorithms, and collecting sample data under various conditions in elastic scene space. Through visual computation, the shooting position of photos was inferred to obtain the approximate position information of users. This method utilizes geographic location information to classify images and combines it with landmarks, natural features, and architectural styles to determine their locations. The experimental results show variations in positioning accuracy among different models, with the most optimal model obtained through training on a large-scale dataset. They also indicate that the positioning error in urban street-based images is relatively small, whereas the positioning effect in outdoor and local scenes, especially in large-scale spatial environments, is limited. This suggests that the location information of users can be effectively determined through the utilization of geographic data, to classify images and incorporate landmarks, natural features, and architectural styles. The study’s experimentation indicates the variation in positioning accuracy among different models, highlighting the significance of training on a large-scale dataset for optimal results. Furthermore, it highlights the contrasting impact on urban street-based images versus outdoor and local scenes in large-scale spatial environments.

A de-texturing model for enhancing the accuracy of centroid positioning

In tasks guided by microvision, extracting centroids is a common method for positioning, which is negatively affected by texture. Here, an attention-related de-texturing model is proposed to eliminate the texture of microparts and preserve accurate edges. A network with an attention module called De-texturing Net is built, in which both the transformer and channel attention modules are included. Considering the importance of texture, the additional factor in loss function is constructed based on the Gram matrix difference between target images and generated images. Results show that De-texturing Net can generate de-texturized images with high Peak Signal to Noise Ratio/SSIM, indicating the similarity between de-texturized and target images. Moreover, for the centroid positioning, the error in de-texturized images is significantly lower than the error in original images. This study helps improve the accuracy of centroid positioning due to the de-texturized images with accurate edges.

A Study on Predicting the Deviation of Jet Trajectory Falling Point under the Influence of Random Wind.

As one of the main external factors affecting the fire extinguishing accuracy of sprinkler systems, it is necessary to analyze and study random wind. However, in practical applications, there is little research on the impact of random wind on sprinkler fire extinguishing points. To address this issue, a new random wind acquisition system was constructed in this paper, and a method for predicting jet trajectory falling points in Random Forest (RF) under the influence of random wind was proposed, and compared with the commonly used prediction model Support Vector Machine (SVM). The method in this article reduces the error in the x direction of the 50 m prediction result from 2.11 m to 1.53 m, the error in the y direction from 0.64 m to 0.6 m, and the total mean absolute error (MAE) from 31.3 to 23.5. Simultaneously, predict the falling points of jet trajectory at different distances under the influence of random wind, to demonstrate the feasibility of the proposed method in practical applications. The experimental results show that the system and method proposed in this article can effectively improve the influence of random wind on the falling points of a jet trajectory. In summary, the image acquisition system and error prediction method proposed in this article have many potential applications in fire extinguishing.

Towards accurate binocular vision of satellites: A cascaded multi-scale pyramid network for stereo matching on satellite imagery

Stereo matching of high-resolution satellite stereo images is a fundamental and challenging task in photogrammetry, geoscience, and remote sensing. Some deep learning-based methods have shown great potential in solving this task recently. However, these methods have not attracted much attention as their accuracy is still unsatisfactory, especially in regions with intractable factors (e.g., occlusions, repetitive patterns, low texture, non-texture, and disparity discontinuities). To tackle these challenging factors and improve disparity accuracy, we propose an end-to-end stereo matching network named Cascaded Multi-Scale Pyramid Network (CMSP-Net). Firstly, to fully utilize the multi-level spatial context information and mitigate the ambiguous matching caused by repetitive patterns, low texture, non-texture, and occlusions, a cost volume pyramid is constructed with multi-scale image features. Then, each cost volume in the pyramid is aggregated with an hourglass structure to exploit the multi-scale context in a single cost volume. Secondly, considering that high-resolution cost volumes contain more detailed information but have small receptive fields, while low-resolution ones have large receptive fields but lack details, we design an Attention-Guided Cost Fusion (AGCF) strategy to fuse pyramid cost volumes, which effectively integrates the advantages of multi-scale cost volumes while mitigating the backfiring caused by semantic inconsistency across different scales. Thirdly, a disparity refinement module is designed to improve overall disparity accuracy further, which incorporates the reconstruction error and edge clues of input satellite stereo images. Extensive experimental results on multiple satellite datasets demonstrate that the proposed method can effectively reduce matching ambiguities and recover sharp disparity discontinuities, achieving clear superiority over existing state-of-the-art methods.

Enhancing voxel-based dosimetry accuracy with an unsupervised deep learning approach for hybrid medical image registration.

Deformable registration is required to generate a time-integrated activity (TIA) map which is essential for voxel-based dosimetry. The conventional iterative registration algorithm using anatomical images (e.g., computed tomography (CT)) could result in registration errors in functional images (e.g., single photon emission computed tomography (SPECT) or positron emission tomography (PET)). Various deep learning-based registration tools have been proposed, but studies specifically focused on the registration of serial hybrid images were not found. In this study, we introduce CoRX-NET, a novel unsupervised deep learning network designed for deformable registration of hybrid medical images. The CoRX-NET structure is based on the Swin-transformer (ST), allowing for the representation of complex spatial connections in images. Its self-attention mechanism aids in the effective exchange and integration of information across diverse image regions. To augment the amalgamation of SPECT and CT features, cross-stitch layers have been integrated into the network. Two different 177 Lu DOTATATE SPECT/CT datasets were acquired at different medical centers. 22 sets from Seoul National University and 14 sets from Sunway Medical Centre are used for training/internal validation and external validation respectively. The CoRX-NET architecture builds upon the ST, enabling the modeling of intricate spatial relationships within images. To further enhance the fusion of SPECT and CT features, cross-stitch layers have been incorporated within the network. The network takes a pair of SPECT/CT images (e.g., fixed and moving images) and generates a deformed SPECT/CT image. The performance of the network was compared with Elastix and TransMorph using L1 loss and structural similarity index measure (SSIM) of CT, SSIM of normalized SPECT, and local normalized cross correlation (LNCC) of SPECT as metrics. The voxel-wise root mean square errors (RMSE) of TIA were compared among the different methods. The ablation study revealed that cross-stitch layers improved SPECT/CT registration performance. The cross-stitch layers notably enhance SSIM (internal validation: 0.9614vs. 0.9653, external validation: 0.9159vs. 0.9189) and LNCC of normalized SPECT images (internal validation: 0.7512vs. 0.7670, external validation: 0.8027vs. 0.8027). CoRX-NET with the cross-stitch layer achieved superior performance metrics compared to Elastix and TransMorph, except for CT SSIM in the external dataset. When qualitatively analyzed for both internal and external validation cases, CoRX-NET consistently demonstrated superior SPECT registration results. In addition, CoRX-NET accomplished SPECT/CT image registration in less than 6s, whereas Elastix required approximately 50s using the same PC's CPU. When employing CoRX-NET, it was observed that the voxel-wise RMSE values for TIA were approximately 27% lower for the kidney and 33% lower for the tumor, compared to when Elastix was used. This study represents a major advancement in achieving precise SPECT/CT registration using an unsupervised deep learning network. It outperforms conventional methods like Elastix and TransMorph, reducing uncertainties in TIA maps for more accurate dose assessments.

An assessment of 2-D and 3-D interest point detectors in volumetric images

Many interest point detectors have been designed so far to work in two dimensional (2-D) images. However, expansion of these detectors into the third dimension for three dimensional (3-D) images can refine their representational power. This paper presents how the Harris corner, LoG filtering-based blob, and salient regions detectors can be expanded to find interest points in volumetric images handling multiple slices collectively. Performances of 2-D and 3-D detector implementations were assessed both qualitatively and quantitatively with value combinations of different parameters using metrics such as F1-score, localization error, and repeatability in binary images of twenty 3-D object models from the Princeton Shape Benchmark (PSB). Computation of F1-score and localization error depended on some manually marked ground truth points, while repeatability measurement was according to the proximity of the detected point sets. The 3-D detectors were evaluated as more successful in capturing distinctive and sparse interest points on 3-D object surfaces in qualitative analyses. Despite having greater computational complexities, most of the 3-D detectors yielded better average F1-score, localization accuracy, and repeatability given uniqueness constraint on the matched points in quantitative analyses. Therefore, the 3-D detectors appear preferable when longer working durations or sparser representations would not constitute any disadvantage.

Chromosome size-dependent polar ejection force impairs mammalian mitotic error correction.

Accurate chromosome segregation requires sister kinetochores to biorient, attaching to opposite spindle poles. To this end, the mammalian kinetochore destabilizes incorrect attachments and stabilizes correct ones, but how it discriminates between these is not yet clear. Here, we test the model that kinetochore tension is the stabilizing cue and ask how chromosome size impacts that model. We live image PtK2 cells, with just 14 chromosomes, widely ranging in size, and find that long chromosomes align at the metaphase plate later than short chromosomes. Enriching for errors and imaging error correction live, we show that long chromosomes exhibit a specific delay in correcting attachments. Using chromokinesin overexpression and laser ablation to perturb polar ejection forces, we find that chromosome size and force on arms determine alignment order. Thus, we propose a model where increased force on long chromosomes can falsely stabilize incorrect attachments, delaying their biorientation. As such, long chromosomes may require compensatory mechanisms for correcting errors to avoid chromosomal instability.

A new ECT image reconstruction algorithm based on Vision transformer (ViT)

Aiming at the problem of low accuracy of ECT image reconstruction, this paper proposes an ECT image reconstruction algorithm based on Vision Transformer (ViT). ViT is a method that applies the Transformer to the field of image classification, which is characterized by strong long-distance dependency learning ability and strong multi-modal fusion ability compared to CNN. This paper fully utilizes the characteristics of ViT to transform the image reconstruction process of ECT into the classification process of ViT. Here, a method is proposed to use the object field distribution as the image classification label. This method converts a two-dimensional image of the object field distribution into a one-dimensional vector, which is the label vector of the image classification. These vectors and their corresponding reconstructed images obtained by the Landweber algorithm form the sample set of ViT. Extracting a large number of flow pattern samples with various shapes and distributions through COMSOL is used as a training set. After training ViT, this network model is used to infer the labels of the predicted flow patterns. After post-processing this label, the corresponding reconstructed image can be obtained. Finally, simulation experiments are conducted, and the experiments results show that the image errors and correlation coefficients of the reconstructed images obtained through the algorithm in this paper are better than those of Tikhonov algorithm, Landweber algorithm and Long Short-Term Memory Network(LSTM). And this algorithm has better resistance to noise interference than Tikhonov algorithm, Landweber algorithm, and LSTM. This also provides a new approach and means for ECT image reconstruction.

Transplantace jater pro hepatocelulární karcinom

Summary: The number of liver transplants for the indication of hepatocellular carcinoma is increasing. We consider the so-called Milan criteria, based on the number and size of the tumour, to be the basic indication framework. However, the original criteria are too restrictive – they prevent access to transplantation for a number of patients who would otherwise be suitable candidates for transplantation. There is therefore an effort to somehow expand these criteria. Decision-making is also burdened by a possible error in imaging – our knowledge of the dimensions and number of nodules is not accurate before transplantation. Later, factors reflecting the biological properties of the tumour were added and it was possible to select patients for transplantation with greater precision. Examples here include the use of alpha-fetoprotein, histological grading, microinvasion into vessels, etc. Also, the response of the tumour to treatment before transplantation can be an important indicator – the effect of locoregional (RFA, MWA, TACE) or even systemic treatment. The importance of pre-transplant administration of immunomodulatory treatment is not yet known. In larger transplant centres, the concept of their own criteria, established on the basis of their own data unique to a given centre, is already beginning to be promoted. The advantage is that it reflects local and specific conditions of the centre. Key words: liver transplantation – hepatocellular carcinoma – Milan criteria – IKEM criteria

TorchQL: A Programming Framework for Integrity Constraints in Machine Learning

Finding errors in machine learning applications requires a thorough exploration of their behavior over data. Existing approaches used by practitioners are often ad-hoc and lack the abstractions needed to scale this process. We present TorchQL, a programming framework to evaluate and improve the correctness of machine learning applications. TorchQL allows users to write queries to specify and check integrity constraints over machine learning models and datasets. It seamlessly integrates relational algebra with functional programming to allow for highly expressive queries using only eight intuitive operators. We evaluate TorchQL on diverse use-cases including finding critical temporal inconsistencies in objects detected across video frames in autonomous driving, finding data imputation errors in time-series medical records, finding data labeling errors in real-world images, and evaluating biases and constraining outputs of language models. Our experiments show that TorchQL enables up to 13x faster query executions than baselines like Pandas and MongoDB, and up to 40% shorter queries than native Python. We also conduct a user study and find that TorchQL is natural enough for developers familiar with Python to specify complex integrity constraints.

Quantum error channels in high energetic photonic systems

In medical applications—such as positron emission tomography (PET)—511 keV photons that experience Compton scattering are studied. We present a consistent framework based on quantum error-correction channels—intensively studied in quantum computing—to fully describe the quantum information-theoretic content of high energetic photons undergoing Compton scattering, characterized by the Klein–Nishina formula in unoriented matter. In this way, we can predict the expected spatial distribution of two or more, pure or mixed, polarization entangled or separable photons. This framework allows us to characterize the accessible and inaccessible information for different parameter ranges. It also answers the question of how to describe successive multi-photon scattering. In addition our formalism provides a complete framework for dealing with single and all multi-partite errors that can occur in the propagation, providing the basis for modeling future dedicated experiments that will then have applications in medicine, such as reducing errors in PET imaging or exploring possibilities for quantum-based diagnostic indicators.

Objective QC for diffusion MRI data: Artefact detection using normative modelling

Abstract Diffusion MRI is a neuroimaging modality used to evaluate brain structure at a microscopic level and can be exploited to map white matter fibre bundles and microstructure in the brain. One common issue is the presence of artefacts, such as acquisition artefacts, physiological artefacts, distortions, or image processing-related artefacts. These may lead to problems with other downstream processes and can bias subsequent analyses. In this work, we use normative modelling to create a semi-automated pipeline for detecting diffusion imaging artefacts and errors by modelling 24 white matter imaging-derived phenotypes from the UK Biobank dataset. The considered features comprised four microstructural features (from models with different complexity such as fractional anisotropy and mean diffusivity from a diffusion tensor model and parameters from neurite orientation, dispersion, and density models), each within six pre-selected white matter tracts of various sizes and geometrical complexity (corpus callosum, bilateral corticospinal tract and uncinate fasciculus and fornix). Our method was compared to two traditional quality control approaches: a visual quality control protocol performed on 500 subjects and quantitative quality control using metrics derived from image pre-processing. The normative modelling framework proves to be comprehensive and efficient in detecting diffusion imaging artefacts arising from various sources (such as susceptibility induced distortions or motion), as well as outliers resulting from inaccurate processing (such as erroneous spatial registrations). This is an important contribution by virtue of this methods’ ability to identify the two problem sources (i) image artefacts and (ii) processing errors, which subsequently allows for a better understanding of our data and informs on inclusion/exclusion criteria of participants.

Reflective optical imaging for scattering medium using chaotic laser.

Optical imaging is a non-invasive imaging technology that utilizes near-infrared light, allows for the image reconstruction of optical properties like diffuse and absorption coefficients within the tissue. A recent trend is to use signal processing techniques or new light sources and expanding its application. We aim to develop the reflective optical imaging using the chaotic correlation technology with chaotic laser and optimize the quality and spatial resolution of reflective optical imaging. Scattering medium was measured using reflective configuration in different inhomogeneous regions to evaluate the performance of the imaging system. The accuracy of the recovered optical properties was investigated. The reconstruction errors of absorption coefficients and geometric centers were analyzed, and the feature metrics of the reconstructed images were evaluated. We showed how chaotic correlation technology can be utilized for information extraction and image reconstruction. This means that a higher signal-to-noise ratio and image reconstruction of inhomogeneous phantoms under different scenarios successfully were achieved. This work highlights that the peak values of correlation of chaotic exhibit smaller reconstruction error and better reconstruction performance in optical imaging compared with reflective optical imaging with the continuous wave laser.

Low-dose CT reconstruction based on high-dimensional partial differential equation projection recovery

We propose a low-dose CT reconstruction method using partial differential equation (PDE) denoising under high-dimensional constraints. The projection data were mapped into a high-dimensional space to construct a high-dimensional representation of the data, which were updated by moving the points in the high-dimensional space. The data were denoised using partial differential equations and the CT image was reconstructed using the FBP algorithm. Compared with those by FBP, PWLS-QM and TGV-WLS methods, the relative root mean square error of the Shepp-Logan image reconstructed by the proposed method were reduced by 68.87%, 50.15% and 27.36%, the structural similarity values were increased by 23.50%, 8.83% and 1.62%, and the feature similarity values were increased by 17.30%, 2.71% and 2.82%, respectively. For clinical image reconstruction, the proposed method, as compared with FBP, PWLS-QM and TGV-WLS methods, resulted in reduction of the relative root mean square error by 42.09%, 31.04% and 21.93%, increased the structural similarity values by 18.33%, 13.45% and 4.63%, and increased the feature similarity values by 3.13%, 1.46% and 1.10%, respectively. The new method can effectively reduce the streak artifacts and noises while maintaining the spatial resolution in reconstructed low-dose CT images.

Research on the grouped orthonormalization method in ghost imaging

Ghost imaging (GI) has found application across diverse fields owing to its distinctive benefits. When employing the rotating ground-glass scheme and utilizing second-order correlation for image reconstruction, the efficiency of imaging is hindered by the multiple sampling of reference patterns. To address this, the orthonormalization method has been employed to enhance image quality and reduce the required number of measurements. Despite its effectiveness, the original orthonormalization method is prone to accumulating imaging noise and errors as the number of measurements increases, leading to a significant degradation in image quality. To overcome this limitation, this paper introduces the grouped orthonormalization method (GO-GI) as an extension of the orthonormalization technique. By adjusting the ‘group size’, this method enables control over the accumulation of errors, resulting in an improvement in image quality. The evaluation of image quality in terms of Contrast-to-Noise demonstrates the significant advantages of the GO-GI method in both simulation and experimental results. This study establishes the GO-GI method as a simple yet practical approach in the realm of GI.

Rotational Convolution Design in Convolutional Neural Networks for Direct 3D Electromagnetic Tomography

The three-dimensional (3D) reconstruction of Electromagnetic Tomography (EMT) is an important task for many applications, such as the non-destructive testing of inner defects in rail systems. Additionally, image reconstruction algorithms utilizing deep learning methods have been verified to be useful in recent years. Therefore, the interpretability of deep learning is a question that is relevant to its application in other areas. This paper proposes an innovative rotational convolution pattern, Conv-P, for convolutional neural network (CNN) image reconstruction in a 3D EMT system. This pattern is based on the projection relationships inherent in tomographic imaging, where each convolution is performed on adjacent projections along the excitation rotation direction. The advantage of this pattern is that it can generate the convolution process by utilizing the 3D structural information from real sensors. To verify the effectiveness of this convolution pattern, we constructed a 3D dual-layer 16-coil EMT model and tested its image reconstruction performance. The results demonstrate that, compared with two common convolution patterns, Conv-P achieves a 4.7% and 4.1% increase in the Image Correlation Coefficient (CC), a 19.8% and 13.1% reduction in the Relative Image Error (IE), a 0.67% and 1.59% increase in the Peak Signal-to-Noise Ratio (PSNR), and a 3.24% and 0.74% increase in the Structural Similarity Index Measure (SSIM).