Trajectory imaging of low-contrast dynamic objects based on frequency-domain analysis.

In order to test the effect of noise on the various parameters of the SAECG, 83 patients underwent three consecutive recordings at different noise levels. The high noise (HN) recordings had a noise level of 0.60-0.74 microV, the intermediate noise (IN) had 0.31-0.59 microV, and the low noise (LN) had < or = 0.30 microV. For the calculation of noise we used the standard deviation of the mean noise of the composite lead high pass filtered at 40 Hz. The recordings were compared using time-domain, frequency-domain, and spectrotemporal analysis. The time-domain parameters of the LN recordings, using 25-Hz, 40-Hz, and 60-Hz high pass cutoffs, were significantly different from those of the HN or IN recordings (P < 0.05). In the frequency-domain analysis, significant differences were found in some of the parameters of the LN compared to the HN. The spectrotemporal analysis of the X and Z leads also showed significant differences among the LN and the other recordings. In the time-domain analysis, both at 40 Hz and 25 Hz, there were more abnormal LN compared to the HN recordings (P < 0.05). In the spectrotemporal analysis, there were significantly more abnormal HN and IN recordings compared to the LN (P < 0.001 and P < 0.01, respectively). Therefore, the level of noise, even within the acceptable range, can significantly affect the SAECG. In the time domain at the lower noise levels the parameters become more abnormal, while the opposite seems to occur in the spectral and the spectrotemporal analysis.

BackgroundIt is difficult to create intuitive methods of controlling prosthetic limbs, often resulting in abandonment. Peripheral nerve interfaces can be used to convert motor intent into commands to a prosthesis. The Extraneural Spatiotemporal Compound Action Potentials Extraction Network (ESCAPE-NET) is a convolutional neural network (CNN) that has previously been demonstrated to be effective at discriminating neural sources in rat sciatic nerves. ESCAPE-NET was designed to operate using data from multi-channel nerve cuff arrays, and use the resulting spatiotemporal signatures to classify individual naturally evoked compound action potentials (nCAPs) based on differing source fascicles. The applicability of this approach to larger and more complex nerves is not well understood. To support future translation to humans, the objective of this study was to characterize the performance of this approach in a computational model of the human median nerve.MethodsUsing a cross-sectional immunohistochemistry image of a human median nerve, a finite-element model was generated and used to simulate extraneural recordings. ESCAPE-NET was used to classify nCAPs based on source location, for varying numbers of sources and noise levels. The performance of ESCAPE-NET was also compared to ResNet-50 and MobileNet-V2 in the context of classifying human nerve cuff data.ResultsClassification accuracy was found to be inversely related to the number of nCAP sources in ESCAPE-NET (3-class: 97.8% ± 0.1%; 10-class: 89.3% ± 5.4% in low-noise conditions, 3-class: 70.3% ± 0.1%; 10-class: 52.5% ± 0.3% in high-noise conditions). ESCAPE-NET overall outperformed both MobileNet-V2 (3-class: 96.5% ± 1.1%; 10-class: 84.9% ± 1.7% in low-noise conditions, 3-class: 86.0% ± 0.6%; 10-class: 41.4% ± 0.9% in high-noise conditions) and ResNet-50 (3-class: 71.2% ± 18.6%; 10-class: 40.1% ± 22.5% in low-noise conditions, 3-class: 81.3% ± 4.4%; 10-class: 31.9% ± 4.4% in high-noise conditions).ConclusionAll three networks were found to learn to differentiate nCAPs from different sources, as evidenced by performance levels well above chance in all cases. ESCAPE-NET was found to have the most robust performance, despite decreasing performance as the number of classes increased, and as noise was varied. These results provide valuable translational guidelines for designing neural interfaces for human use.

The ground penetrating radar (GPR) has demonstrated good potential for the remote imaging of surface-laid or shallow-buried landmine-like objects (typically ~ 5-30 cm) and its use is currently receiving much attention. It has turned out to be a promising alternative technology for low dielectric contrast objects, a difficult detection situation that is often encountered in practise (e.g. detection of plastic mines in dry or sandy soils environment). This paper examines numerically the imaging of buried objects using ultrawide-band (UWB) time domain radar. We develop a simplified model to characterize the system, air-ground-buried targets-antennas and simulate the electromagnetic wave propagation and scattering at a bandwidth of 0.5-2.5 Ghz. All the elements need to be modelled simultaneously in order to obtain an accurate estimation of our radar performance and surface response to the incoming radar pulses. The final goal is to improve the detection rate of plastic antipersonnel mines, reducing the false alarm level. We show some simulated results in 2D and 3D assuming plane waves and afterwards we introduce a model for our GPR antennas to characterize the real source.

A feasibility study is presented for acoustic tomographic imaging of arbitrarily shaped objects in range-independent oceanic waveguides. One of the main obstacles in the efficient use of tomographic imaging is that tomographic systems usually allow for a number of different independent solutions which satisfy the same systems of equations. This adverse property of acoustic tomography is especially pronounced in underdetermined systems characterized by null spaces of tomographic matrices which relate measured data to an unknown sound-speed distribution. In fact, acoustic tomography is usually underdetermined and the null space problem becomes critical in many applications. Acoustic tomography in many instances yields blurred images precluding its efficient use for mapping underwater objects in oceanic waveguides. How to improve the resolving power of tomographic imagery in waveguides is shown. A new technique based on the sliding window approach is suggested which uses constrained traveltime inversion. The method will be tested on mapping opaque and low-contrast acoustic objects placed in range-independent waveguides. The method is fast since it does not directly invert tomographic matrices of large rank but instead carries out summation over nonzero elements of tomographic matrices. It can be proved that constrained inversion compensates for limited angular aperture recording and thus can be applied to practical problems with incomplete angular aperture. This work is limited to processing of transmitted signals. A separate study will be devoted to the analysis of reflection data.

One of the signs of early breast cancer is the presence of microcalcification clusters in mammograms. Automated diagnostic aids for detection of breast cancer in mammograms must determine the presence of microcalcification clusters with high enough accuracy to serve as a clinical tool. To do so, the automated system has to process the digitized mammogram with at least three different types of algorithms. The first is segmentation of individual small candidate structures that are separated from the background tissue. The second is extraction of quantitative features from each candidate structure, to represent size, shape, texture, contrast, etc. The third is classification of candidate structures as microcalcification or other tissue, using the feature values, based on automated training against ground truth provided by clinical experts. An optional fourth type of algorithm, enhancement, may also be used as a preprocessing step, depending on the image quality, to reduce the noise level of the image or to increase the contrast of candidate objects before segmentation is applied. In this multistage process, the role of segmentation is essential because the delineation of each candidate object dictates the extracted features, and subsequently their classification. This chapter focuses on segmentation algorithms that are especially suited for segmenting small, low-contrast, arbitrarily shaped objects such as microcalcifications in mammograms. Segmentation algorithms used to separate objects of interest from the image background are primarily based on three different approaches. Some algorithms determine the pixels of the object by comparing their brightness values to either a global or local threshold. The second type of algorithm first determines the edge pixels of the object using a gradient operator, then an edge-linking technique forms a closed contour and labels all pixels within the edge contour. The third approach is to grow a region of contiguous pixels starting from a seed pixel selected inside the object. The segmentation of microcalcifications in mammograms is especially demanding because each microcalcification is a very small, low-contrast object embedded in a nonuniform background that may have significant spatial high-frequency components. An edge detector applied to a mammogram finds edges throughout the image and may also overlook the edges of some microcalcifications due to their low contrast. Segmentation of small, low-contrast objects such as microcalcifications cannot effectively be based on conventional edge detection, especially because linking edge pixels obtained on a mammogram is not practical. Most algorithms reported for segmentation of microcalcifications are based either on thresholding or region growing, with or without enhancement. The segmentation algorithm suggested in Refs. 1 and 2 compares each pixel to a local threshold computed within a square neighborhood around the pixel. The size of the neighborhood is selected by the user, and the threshold is set as the mean value of pixels in the neighborhood plus their rms value multiplied by a selected coefficient. This threshold represents the highest value that the background pixels in that neighborhood are expected to have. Pixels that have values greater than the threshold are connected to form the segmented region.

Welding industry is one of the leaders in the field of investments in the development of technologies and innovations aimed at ensuring functional cost-effectiveness, improving safety and speed of movement in the digital age. Ensuring the reliability and safety of the railway tracks requires the development of new scientific approaches to the use of welding technologies in the construction and operation of a high-speed railway. The main criterion for evaluating efficiency of the new approaches is the quality of the resulting welded joints and the productivity of the welding processes used. Rail connections should be made in such a way from a technical point of view that there are no restrictions during operation, the service life of the rails is not reduced, the percentage of defects does not increase, and the noise level is within the normal limits. The purpose of the study is to substantiate the cost-effectiveness and safety of the introduction of new contact welding technologies in Russia. Technical and economic analysis of welding methods is given, including using the rail welding machines — a new generation of equipment for the contact butt welding of high-strength rails during the construction and reconstruction of a high-speed railway. The Up-to-date technologies are based on the process of contact welding by pulsating reflow, on the one hand ensuring high productivity of the process, and on the other — reducing the influence of the human factor on it. This is achieved through a set of options for controlling the welding process, with online monitoring of the operational characteristics, with the software for storage and the possibility of using robotic welding in the future in the field of digital doubles and processing archived data based on the artificial intelligence as part of the implementation of Smart City projects. Feasibility study is also presented concerning the efficiency of the introduction of contact welding technology by the pulsed reflow on high-speed railway projects of Russia based on reducing the rail joint failures and improving the characteristics of the welding process and safe operation.

Purpose: MTF of iterative reconstruction (IR) CT is known to be contrast dependent, which is of critical importance for predicting signal response of lesions with different contrast. Previous studies attempting to characterize the contrast dependant MTF of IR CT were carried out using multi contrast phantoms, which is of limited effect in revealing the range of MTF change between extremely high and low contrast objects. This study presents a novel technique for assessing the range of MTF changes and compares the measurement results in four IR CT. Methods: An AAPM phantom was scanned with 4 different CT systems (Discovery 750, GE, Somatom Definition, Siemens, Ingenuity, Philips, and Aquillion One, Toshiba) at same dose level (12.5mGy), and was reconstructed using FBP and IR techniques (ASIR and VEO in GE, Safire in Siemens, iDose4 in Philips, and AIDR in Toshiba). Total MTF was measured using a small bead in the phantom to characterize frequency response for high contrast objects. In addition, the algorithmic MTF was derived from noise power spectrum (NPS) using uniform background images in order to evaluate the frequency response for very low contrast objects. Results: The algorithmic MTF 50 of FBP showed higher responses up to 16.5% than total MTF 50 in all 4 CT systems, which reflected the decoupling of focal spot blurring effect. The algorithmic MTF 50 ofIR recons was lower that total MTF 50 in2 vendors(GE, Siemens) showing up to 48% reduction of MTF 50 in VEO, whereas IR recons from the other 2 vendors showed no marked reduction of algorithmic MTF 50. Conclusion: Our study revealed the degrading trends of MTF for low contrast objects in some IR technique. This characteristic needs to be considered in interpreting signal patterns of low contrast lesions in IR CT. The research was supported by the Converging Research Center Program through the Ministry of Education, Science and Technology (2012K001498).

Noisy inputs pose significant challenges for reinforcement learning (RL) agents navigating real-world environments. While animals demonstrate robust spatial learning under dynamic conditions, the mechanisms underlying this resilience remain understudied in RL frameworks. This paper introduces a novel comparative analysis of predecessor feature (PF) and successor feature (SF) algorithms under controlled noise conditions, revealing several insights. Our key innovation lies in demonstrating that SF algorithms achieve superior noise resilience compared to traditional approaches, with cumulative rewards of 2216.88±3.83 (mean ± SEM), even under high noise conditions (σ=0.5) in one-dimensional environments, while Q learning achieves only 19.22±0.57. In two-dimensional environments, we discover an unprecedented nonlinear relationship between noise level and algorithm performance, with SF showing optimal performance at moderate noise levels (σ=0.25), achieving cumulative rewards of 2886.03±1.63 compared to 2798.16±3.54 for Q learning. The λ parameter in PF learning is a significant factor, with λ=0.7 consistently achieving higher λ values under most noise conditions. These findings bridge computational neuroscience and RL, offering practical insights for developing noise-resistant learning systems. Our results have direct applications in robotics, autonomous navigation, and sensor-based AI systems, particularly in environments with inherent observational uncertainty.

Purpose: Luggage X-rays suffer from low contrast, material overlap, and noise; dual-energy imaging reduces ambiguity but creates colour biases that impair segmentation. This study aimed to (1) employ connotative fusion by embedding realistic detonator patches into real X-rays to simulate threats and enhance unattended detection without requiring ground-truth labels; (2) thoroughly evaluate fusion techniques in terms of balancing image quality, information content, contrast, and the preservation of meaningful features. Methods: A total of 1000 X-ray luggage images and 150 detonator images were used for fusion experiments based on deep learning, transform-based, and feature-driven methods. The proposed approach does not need ground truth supervision. Deep learning fusion techniques, including VGG, FusionNet, and AttentionFuse, enable the dynamic selection and combination of features from multiple input images. The transform-based fusion methods convert input images into different domains using mathematical transforms to enhance fine structures. The Nonsubsampled Contourlet Transform (NSCT), Curvelet Transform, and Laplacian Pyramid (LP) are employed. Feature-driven image fusion methods combine meaningful representations for easier interpretation. Singular Value Decomposition (SVD), Principal Component Analysis (PCA), Random Forest (RF), and Local Binary Pattern (LBP) are used to capture and compare texture details across source images. Entropy (EN), Standard Deviation (SD), and Average Gradient (AG) assess factors such as spatial resolution, contrast preservation, and information retention and are used to evaluate the performance of the analysed methods. Results: The results highlight the strengths and limitations of the evaluated techniques, demonstrating their effectiveness in producing sharpened fused X-ray images with clearly emphasized targets and enhanced structural details. Conclusions: The Laplacian Pyramid fusion method emerges as the most versatile choice for applications demanding a balanced trade-off. This is evidenced by its overall multi-criteria balance, supported by a composite (geometric mean) score on normalised metrics. It consistently achieves high performance across all evaluated metrics, making it reliable for detecting concealed threats under diverse imaging conditions.

Ss performed in a button-pushing task in which 0.75 sec of either 95-dB SPL or 30-dB SPL noise was delivered contingent on each response for a RRLRL pattern of responses. Ss who received high-level noise in their first exposure to the task generally learned to avoid the noise. About half (13) of the Ss who received low-level noise in their first exposure to the task learned to avoid the noise, while the other half (12) learned to deliver the noise on every response. Eight of the 12 Ss who learned to deliver low noise continued to deliver noise for every response when the noise level was increased to 95 dB. There were no consistent differences in the amount of time required to learn the pattern in low-noise and high-noise conditions.

Accurate segmentation in medical imaging is essential for disease diagnosis and monitoring, particularly in lung imaging using proton and hyperpolarized gas MRI. However, image degradation due to noise and artifacts—especially in hyperpolarized gas MRI, where scans are acquired during breath-holds—poses challenges for conventional segmentation algorithms. This study evaluates the robustness of deep learning segmentation models under varying Gaussian noise levels, comparing traditional convolutional neural networks (CNNs) with modern Vision Transformer (ViT)-based models. Using a dataset of proton and hyperpolarized gas MRI slices from 56 participants, we trained and tested Feature Pyramid Network (FPN) and U-Net architectures with both CNN (VGG16, VGG19, ResNet152) and ViT (MiT-B0, B3, B5) backbones. Results showed that ViT-based models, particularly those using the SegFormer backbone, consistently outperformed CNN-based counterparts across all metrics and noise levels. The performance gap was especially pronounced in high-noise conditions, where transformer models retained higher Dice scores and lower boundary errors. These findings highlight the potential of ViT-based architectures for deployment in clinically realistic, low-SNR environments such as hyperpolarized gas MRI, where segmentation reliability is critical.

The main result of this work is the creation of a methodology for determining the coordinates, velocities and trajectories of spatial movements of moving objects by means of kinematic design and the development of principle schemes for the optimal placement of radar search equipment for the successful detection of enemy UAVs. Its purpose is to increase the defense capability of military units, populated areas of Ukraine and their infrastructure by resisting attacks by attack and sabotage reconnaissance unmanned aerial vehicles (UAVs). The protection scheme is based on the determination of coordinates, movement speeds and trajectories of spatial movements of enemy aircraft by means of kinematic design. The established coordinates of enemy UAVs are reported to command posts equipped with liquidator drones (kamikaze) and small arms or to control points of anti-aircraft missile systems (SAMS), which eliminate these aircraft. The proposed improvement of the widespread method of radar search for moving objects is based on the introduction of additional search radar equipment and mathematical apparatus for precise calculations of coordinates and parameters of spatial movements of moving objects by means of kinematic design. Two paired radar search systems, equipped with modern high-speed computing equipment with appropriate software, are able to "screen out" false return electromagnetic signals. This makes it possible to clearly single out the wanted moving object and subsequently track its movement in space. Corresponding mathematical dependencies are proposed for calculating the distance of a moving object to radar stations, its coordinates in the introduced orthogonal coordinate system, speed and acceleration of movement, that is, instantaneous coordinates and parameters of spatial movements of moving objects. This makes it possible to increase the effectiveness of detecting enemy attack and sabotage-reconnaissance unmanned aerial vehicles, including Iranian-made drones of the Sahed-136 model.

Stability of a car's trajectory of motion is the ability of a car to retain its required state of motion, under different conditions of motion. Depending on different states of motion, cars can lose direction of motion instability when there is a lateral force, due to vehicle speed, due to turning or braking the road with different grip coefficients on the sides of the wheel. In such complex moving conditions, cars need to keep a stable trajectory of movement under the driver's control. The paper examines the stability of the car's trajectory in linear motion conditions with the influence of lateral forces and the velocity of the vehicle, these two factors act to change the state of movement of the vehicle, which significantly affects the safety of the vehicle's movement. The authors looked at the effect of vehicle velocity and lateral forces on vehicle transverse dynamics: lateral forces caused by horizontal winds, inclination of the road surface, speed of movement of the vehicle acting on the entire body create centrifugal inertial forces; this force acts to cause the tires to deform. At this point the vehicle is shifted from its original trajectory. In the research content, the authors built a lateral Dynamics Model, in which the influencing factors are set as inputs to the model including vehicle velocity and transverse wind forces, output is the rotation angle of the body, transverse acceleration and horizontal displacement of the vehicle. The simulation results showed that the effect of vehicle speed was relatively large: lateral displacement increased by 228 % at a speed of 120 km/h. Therefore, when driving on the highway, there should always be driver control to ensure that the car does not deviate from the lane

Virtual trajectories of single-joint movements performed under two basic strategies

Many studies on human motor control have examined how humans move the arms. A well known phenomenon known as the Fitts' law describes the trade-off between the speed of motion and its accuracy. In robotics, many studies were performed, and different control methods were proposed, for robots working in well structured environments. Nevertheless, the cooperation between humans and robots remains a challenging task that is highly relevant for robots aiming to work together with humans in non-structured environments. In this paper we propose a novel method for on-line adaptation of robotic trajectories, where humans and robots are autonomous agents coupled through physical interaction, for example through manipulated object. Within the proposed framework, the robot adapts to the human motion through the sensory feedback by taking into account the Fitts' law where the user specific behavior is estimated using a recursive least square updates. The movement trajectories are represented by the Dynamic Movement Primitives, where the adaptation relies on the Iterative Learning Controller framework. The proposed approach was evaluated by a cooperative human robot arm-reaching task on a plane. We tested the accuracy and efficiency of the proposed method, showed its relevance and demonstrated that the proposed approach fully adapts to the human motor control and maintains the desired accuracy of the movement.