Pētniecība
Publikācijas
729 publikācijas
- 2023
Unintentional drifts in performance during one-hand and two-hand finger force production(atveras jaunā logā)
Valters Abolins, Juris Ormanis, Mark Latash
Experimental Brain Research, Sēj. 241 (3), lpp. 699-712, Springer Nature
Kopsavilkums
We explored the phenomena of force drifts and unintentional finger force production (enslaving), and their dependence on visual feedback. Predictions have been drawn based on the theory of control with spatial referent coordinates for condition with feedback on instructed (master) finger force, enslaved finger force, and total force for one-hand and two-hand tasks. Subjects produced force under the different feedback conditions without their knowledge. No feedback condition was also used for the one-hand tasks. Overall, feedback of master finger force led to an increase in the enslaved force, feedback on the slave finger force led to a drop in the master force, feedback on the total force led to balanced drifts in the master force down and enslaved force up, and under the no-feedback condition, master and total force drifted down with large variability in the enslaved force drifts. The patterns were the same in both hands in the two-hand tasks when feedback was provided on the forces of one hand only (without subject’s knowledge). The index of enslaving always drifted toward higher values. We interpret the findings as reflecting three main factors: drifts in the referent coordinates toward actual finger coordinates, spread of cortical excitation over representations of the fingers, and robust sharing of referent coordinates between the two hands in bimanual tasks. The large consistent drifts in enslaving toward higher values have to be considered in studies of multi-finger synergies.
Indeksēts Scopus / WoS · Scopus procentile: 38
- 2023
Discrete Denoising Diffusion Approach to Integer Factorization(atveras jaunā logā)
Kārlis Freivalds, Emīls Ozoliņš, Guntis Bārzdiņš
32nd International Conference on Artificial Neural Networks, Sēj. 14254 (1), lpp. 123-134, Lecture Notes in Computer Science (LNCS)
Kopsavilkums
Integer factorization is a famous computational problem unknown whether being solvable in the polynomial time. With the rise of deep neural networks, it is interesting whether they can facilitate faster factorization. We present an approach to factorization utilizing deep neural networks and discrete denoising diffusion that works by iteratively correcting errors in a partially-correct solution. To this end, we develop a new seq2seq neural network architecture, employ relaxed categorical distribution and adapt the reverse diffusion process to cope better with inaccuracies in the denoising step. The approach is able to find factors for integers of up to 56 bits long. Our analysis indicates that investment in training leads to an exponential decrease of sampling steps required at inference to achieve a given success rate, thus counteracting an exponential run-time increase depending on the bit-length. https://arxiv.org/abs/2309.05295
Indeksēts Scopus / WoS · Scopus procentile: 51
- 2023
Augmenting a Pretrained Object Detection Model with Planar Pose Estimation Capability(atveras jaunā logā)
Andris Lapins, Janis Arents and Modris Greitans
Automatic Control and Computer Sciences, Sēj. 57 (5), lpp. 459-468, Pleiades Publishing
Kopsavilkums
This paper presents a 2D pose estimation solution to the bin-picking problem for robotic grasping systems. By extending a pretrained object detection model, namely DETR, with pose and visibility prediction heads we obtain classification, center, 2D rotation and occlusion scores for every detected object. The augmented model is trained and evaluated on synthetically generated images representing the real environment for faster and more flexible acquisition of data. The results show an average angle error of 3.23 deg for cylindrical and cuboid shape objects.
Indeksēts Scopus / WoS · Scopus procentile: 28 · SNIP: 0.47
- 2023
A. Study of Ultrasonic Guided Wave Propagation in Bone Composite Structures for Revealing Osteoporosis Diagnostic Indicators(atveras jaunā logā)
Glushkov E, Glushkova N, Ermolenko O, Tatarinov A
Materials, 2023, Sēj. 16 (19), lpp. 6179
Kopsavilkums
Tubular bones are layered waveguide structures composed of soft tissue, cortical and porous bone tissue, and bone marrow. Ultrasound diagnostics of such biocomposites are based on the guided wave excitation and registration by piezoelectric transducers applied to the waveguide surface. Meanwhile, the upper sublayers shield the diseased interior, creating difficulties in extracting information about its weakening from the surface signals. To overcome these difficulties, we exploit the advantages of the Green’s matrix-based approach and adopt the methods and algorithms developed for the guided wave structural health monitoring of industrial composites. Based on the computer models implementing this approach and experimental measurements performed on bone phantoms, we analyze the feasibility of using different wave characteristics to detect hidden diagnostic signs of developing osteoporosis. It is shown that, despite the poor excitability of the most useful modes associated with the diseased inner layers, the use of the improved matrix pencil method combined with objective functions based on the Green’s matrix allows for effective monitoring of changes in the elastic moduli of the deeper sublayers. We also note the sensitivity and monotonic dependence of the resonance response frequencies on the degradation of elastic properties, making them a promising indicator for osteoporosis diagnostics.
Indeksēts Scopus / WoS · Scopus procentile: 70
- 2023
Evaluation of Factors-of-Interest in Bone Mimicking Models Based on DFT Analysis of Ultrasonic Signal(atveras jaunā logā)
Sisojevs A, Tatarinov A, Chaplinska A
In Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods (ICPRAM 2023), lpp. 914-919
Kopsavilkums
Bone fragility in osteoporosis is associated with a decrease in the thickness of the cortical layer CTh in long bones and the development of internal porosity P. In the present work, an attempt was made to predict the factors-of-interest CTh and P based on the pattern recognition approach, where DFT analysis was applied to ultrasonic signals in surface transmission through a soft tissue layer. Compact bone was modeled with PMMA plates with gradual changes in CTh from 2 to 6 mm, and internal porosity P was created by drilling where the thickness of the porous layer P varied from 0 to 100% of CTh. The estimation method was based on a statistical analysis of the magnitude of the DFT spectrum of the ultrasonic signals. Decision rules were mathematical criteria calculated as ratios between the envelope functions of the magnitudes. Each of the objects was chosen in turn as a test object, while other specimens composed the training set. The results of the experiments showed the potential effectiveness of the CTh and P prediction, while additional physical parameters may be used as decision rules to improve the reliability of the diagnosis.
Indeksēts Scopus / WoS
- 2023
Three-Dimensional Imaging in Agriculture: Challenges and Advancements in the Phenotyping of Japanese Quinces in Latvia(atveras jaunā logā)
Edīte Kaufmane, Edgars Edelmers, Kaspars Sudars, Ivars Namatēvs, Arturs Nikulins, Sarmīte Strautiņa, Ieva Kalniņa, Astile Peter
Horticulturae, Sēj. 9 (12), lpp. 16, MDPI
Kopsavilkums
This study presents an innovative approach to fruit measurement using 3D imaging, focusing on Japanese quince (Chaenomeles japonica) cultivated in Latvia. The research consisted of two phases: manual measurements of fruit parameters (length and width) using a calliper and 3D imaging using an algorithm based on k-nearest neighbors (k-NN), the ingeniously designed “Imaginary Square” method, and object projection analysis. Our results revealed discrepancies between manual measurements and 3D imaging data, highlighting challenges in the precision and accuracy of 3D imaging techniques. The study identified two primary constraints: variability in fruit positioning on the scanning platform and difficulties in distinguishing individual fruits in close proximity. These limitations underscore the need for improved algorithmic capabilities to handle diverse spatial orientations and proximities. Our findings emphasize the importance of refining 3D scanning techniques for better reliability and accuracy in agricultural applications. Enhancements in image processing, depth perception algorithms, and machine learning models are crucial for effective implementation in diverse agricultural scenarios. This research not only contributes to the scientific understanding of 3D imaging in horticulture but also underscores its potential and limitations in advancing sustainable and productive farming practices.
Indeksēts Scopus / WoS · Scopus procentile: 85 · SNIP: 0.969
- 2023
Human Body as a Signal Transmission Medium for Body-Coupled Communication: Galvanic-Mode Models(atveras jaunā logā)
V. Aristov, A. Elsts.
Electronics, Sēj. 12 (21), lpp. 4550, MDPI
Kopsavilkums
Signal propagation models play a fundamental role in radio frequency communication research. However, emerging communication methods, such as body-coupled communication (BCC), require the creation of new models. In this paper, we introduce mathematical models that approximate the human body as an electrical circuit, as well as linear regression- and random forest-based predictive models that infer the expected signal loss from its frequency, measurement point locations, and body parameters. The results demonstrate a close correspondence between the amplitude-frequency response (AFR) predicted by the electrical circuit models and the experimental data gathered from volunteers. The accuracy of our predictive models was assessed by using their root mean square errors (RMSE), ranging from 1.5 to 7 dB depending on the signal frequency within the 0.05 to 20 MHz range. These results allow researchers and engineers to simulate and forecast the expected signal loss within BCC systems during their design phase.
Scopus procentile: 68
- 2023
Adaptive Signal-to-Noise Ratio Indicator for Wearable Bioimpedance Monitoring(atveras jaunā logā)
D. Lapsa, R. Janeliukstis, A. Elsts
Sensors, Sēj. 23 (20), lpp. 8532, MDPI
Kopsavilkums
Bioimpedance monitoring is an increasingly important non-invasive technique for assessing physiological parameters such as body composition, hydration levels, heart rate, and breathing. However, sensor signals obtained from real-world experimental conditions invariably contain noise, which can significantly degrade the reliability of the derived quantities. Therefore, it is crucial to evaluate the quality of measured signals to ensure accurate physiological parameter values. In this study, we present a novel wrist-worn wearable device for bioimpedance monitoring, and propose a method for estimating signal quality for sensor signals obtained on the device. The method is based on the continuous wavelet transform of the measured signal, identification of wavelet ridges, and assessment of their energy weighted by the ridge duration. We validate the algorithm using a small-scale experimental study with the wearable device, and explore the effects of variables such as window size and different skin/electrode coupling agents on signal quality and repeatability. In comparison with traditional wavelet-based signal denoising, the proposed method is more adaptive and achieves a comparable signal-to-noise ratio.
Indeksēts Scopus / WoS · Scopus procentile: 87
- 2023
ETAM: Next Generation Event Timer for Picosecond-Precision Time and Amplitude Measurements(atveras jaunā logā)
A. Kalinovskis, V. Stepanovs, A. Ancans, A. Elsts
Sensors, Sēj. 23 (14), lpp. 6380, MDPI
Kopsavilkums
Riga Event Timers have the ability to measure the interval between events with high resolution, on the order of picoseconds. However, they have several drawbacks, such as sensitivity to environmental temperature changes and an inability to capture the amplitude of the events. In this work, we present the ETAM: a next generation Event Timer. Its innovative features include adaptive correction of measurement errors based on an internal temperature sensor, and integrated peak-detector circuit to determine the amplitude of nanosecond-duration pulses. Evaluation shows that the ETAM has high thermal stability with a root mean square error (RMSE) of <3 ps in a temperature range between 0 and +40 °C, and accurate event amplitude measurement capability, with <2.3 mV RMSE in the 100–1000 mV range. These improvements allow the ETAM to be used in satellite laser ranging, optical time-domain reflectometry, and other field applications that require temperature- and amplitude-based time correction in addition to high robustness, performance, and stability.
Scopus procentile: 87
- 2023
Demo: Audio Streaming Via Body-Coupled Communication
V. Medvedevs, J. Ormanis, D. Tiščenko, A. Elsts
EWSN '23, ACM
Kopsavilkums
This demo presents an implementation of Body-Coupled Communication (BCC) technology, which enables the use of the human body as a communication medium by passing a low-voltage, low-frequency (less than 20 MHz) modulated electrical field through it. The BCC Demo setup is based on two software-defined radios and demonstrates real-time, standard-quality audio streaming over a communication link formed by the human body. This demo could serve as an example and inspiration to build BCC systems that have the potential to enable more secure and energy-efficient body-area sensor networks and data transfer applications.
- 2023
Denoising Diffusion Algorithm for Single Image Inplaine Super-Resolution in CBCT Scans of the Mandible(atveras jaunā logā)
Ivars Namatevs, Kaspars Sudars, Arturs Nikulins, Anda Slaidina, Laura Neimane, Oskars Radzins, Edgars Edelmers
2023 IEEE 64th International Scientific Conference on Information Technology and Management Science of Riga Technical University (ITMS), IEEE
Kopsavilkums
Deep Learning models are currently the cornerstone of artificial intelligence in medical imaging. The performance of Deep Learning in medical imaging is significantly influenced by the amount and quality of training data. Diffusion models have recently attracted the attention of the computer vision community as they enable photorealistic synthetic image-to-image translation. Previous attempts to use diffusion models for super-resolution imaging have produced satisfactory high-resolution images from low-resolution inputs. However, the drawback is the slow speed of inference, which severely hinders practical applications in medicine. To speed up inference and further improve performance, we propose an accelerated algorithm based on denoising diffusion probability modelling approach for medical image super-resolution. Instead of sampling from pure Gaussian noise, the intermediate distributions of noisy low- and high-resolution images are compared and used to generate super-resolution images. Our proposed algorithm is used to convert low-resolution panoramic X-ray images from Cone-beam Computed Tomography scans of the mandible into high-resolution images for the identification of osteoporosis.
Indeksēts Scopus / WoS
- 2023
Natural Language Conditioned Planning of Complex Robotics Tasks(atveras jaunā logā)
Toms Eduards Zinars, Oskars Vismanis, Peteris Racinskis, Janis Arents, Modris Greitans
Advancing Edge Artificial Intelligence, lpp. 131-152, River Publishers Series in Communications and Networking
Kopsavilkums
Mapping the environment is a powerful technique for enabling autonomy through localization and planning in robotics. This article seeks to provide a global overview of actionable map construction in robotics, outlining the basic problems, introducing techniques for overcoming them, and directing the reader toward established research covering these problem and solution domains in more detail. Multiple levels of abstraction are covered in a non-exhaustive vertical slice, starting with the fundamental problem of constructing metric occupancy grids with Simultaneous Mapping and Localization techniques. On top of these, topological meshes and semantic maps are reviewed, and a comparison is drawn between multiple representation formats. Furthermore, the datasets and metrics used in performance benchmarks are discussed, as are the challenges faced in some domains that deviate from typical laboratory conditions. Finally, recent advances in robot control without explicit map construction are touched upon.
- 2023
Wired Communication Protocol for Body Sensor Network: A Solution for Sensor Data Acquisition(atveras jaunā logā)
Armands Ancans
IWoEDI'2023
Kopsavilkums
Body sensor networks typically rely on wireless communication technologies, but in large and dense topologies, challenges with material efficiency, cost-effectiveness, and coex-istence can arise. Wired communication has the potential to address these issues while preserving wearability and movement freedom, yet wired networks for body sensor networks have been overlooked and under-studied. In this extended abstract, an architecture for a low-power wired body sensor network and an optimized communication protocol for a large number of sensor nodes is proposed. And a proof-of-concept experimental device that demonstrates the feasibility and suitability of our approach for real-time applications is presented.
- 2023
Digital Twins for CCAM Applications - the Case of AUGMENTED CCAM and Beyond(atveras jaunā logā)
Evangelos Bekiaris, Maria Gkemou, Ana Martínez, Amr Elkenawy, Roberts Kadikis, Kaspars Ozols, and others
Relstat2023, TSI
Kopsavilkums
Digital Twins (DT) are increasingly penetrating all sectors of digitized work, with the transport sector and especially the domain of Cooperative Connected and Automated Mobility (CCAM) following rapidly. The current manuscript provides an insight in the DT for transport and CCAM and presents the development approach to be followed in the AUGMENTED CCAM EU funded project, that encompasses in multiple ways and levels the use of DT for the Physical and Digital Infrastructure (PDI) enabled solutions that will be deployed in the test sites of Latvia, Spain and France to assist with Operational Design Domain (ODD) extension of Connected and Automated Vehicles (CAVs).
- 2023
The IMOCO4.E reference framework for intelligent motion control systems(atveras jaunā logā)
Sajid Mohamed, Gijs van der Veen, Hans Kuppens, Matias Vierimaa, Tassos Kanellos, Henry Stoutjesdijk, Riccardo Masiero, Kalle Määttä, Jan Wytze van der Weit, Gabriel Ribeiro, Ansgar Bergmann, Davide Colombo, Javier Arenas, Alfie Keary, Martin Goubej, Benjamin Rouxel, Pekka Kilpeläinen, Roberts Kadikis, Mikel Armendia, Petr Blaha, Joep Stokkermans, Martin Čech, Arend-Jan Beltman
IEEE International Conference on Emerging Technologies and Factory Automation, lpp. 1-8, IEEE
Kopsavilkums
Intelligent motion control is integral to modern cyber-physical systems. However, smart integration of intelligent motion control with commercial and industrial systems requires domain expertise, industrial 'know-how' of the production processes, and resilient adaptation for the various engineering phases. The challenge is amplified with the adoption of advanced digital twin approaches, big data and artificial intelligence in the various industrial domains. This paper proposes the IMOCO4.E reference framework for the smart integration of intelligent motion control with commercial platforms (e.g. from SMEs) and industrial systems. The IMOCO4.E reference framework brings together the architecture, data management, artificial intelligence and digital twin viewpoints from the industrial users of the large-scale 'Intelligent Motion Control under Industry4.E' (IMOCO4.E) consortium. The framework envisions a generic platform for designing, developing, and implementing novice and complex motion-controlled industrial systems. Refinements and instantiations of the framework for the IMOCO4.E industrial cases validate the framework's applicability for various industrial domains throughout the engineering phases and under different constraints imposed on the industrial cases. © 2023 IEEE.
- 2023
Synthetic Image Generation With a Fine-Tuned Latent Diffusion Model for Organ on Chip Cell Image Classification(atveras jaunā logā)
Maksims Ivanovs, Laura Leja, Kārlis Zviedris, Roberts Rimsa, Karina Narbute, Valerija Movcana, Felikss Rumnieks, Arnis Strods, Kevin Gillois, Gatis Mozolevskis, Arturs Abols, Roberts Kadikis
Signal Processing - Algorithms, Architectures, Arrangements, and Applications Conference Proceedings, SPA, lpp. 148-153, IEEE
Kopsavilkums
Augmentation of the datasets of authentic microscopy images with synthetic images is a promising solution to the problem of the limited availability of biomedical data for training deep neural network (DNN) based classifiers. In the present study, we use a text-to-image latent stable diffusion model fine-tuned by means of low-rank adaptation (LoRA) to augment a small dataset of the images of organ on chip cells. While the resulting synthetic images appear quite similar to the authentic images on which the low-rank adaptation was performed, we find that neither training the EfficientNetB7 DNN model solely on the synthetic data nor augmentation of the real-world dataset with different proportions (10, 25, 50, and 75 percent) of these data leads to the improvement of the accuracy of the model. The findings of our study suggest that a further exploration of the low-rank adaptation options is needed to fully use the capacity of latent diffusion models for the synthesis of biomedical images.
- 2023
Modified SERVPERF vs Modified SERVSAT: Comparison of Methods for Measuring the Educational Services Quality(atveras jaunā logā)
Ruslana Moskotina, Mykola Sydorov, Juris Breidaks
Sociological Studios, Sēj. 1 (22), lpp. 65-75
Kopsavilkums
The article discusses two methods designed to measure the quality of educational services are included in the UNiDOS monitoring survey. There are modified SERVPERF and modified SERVSAT. They consist of the same questions that characterize various aspects of the educational process but have different scales. The modified SERVSAT involves direct measurement of satisfaction with educational process. The modified SERVPERF measures satisfaction indirectly. It compares respondents’ expectations about educational process and the reality. We tried to find out whether modified SERVPERF and modified SERVSAT can be considered equivalent. In general, these methods give the same results. They have the same factor structure. The questions grouped into five dimensions: Tangibles, Reliability, Assurance, Responsiveness, Empathy. But there are also differences that do not allow us to definitively claim that the methods are equivalent. Comparing the means shows that the scale of the modified SERVPERF is slightly shifted to the left relatively to the scale of the modified SERVSAT. Analysis of the percentage of respondents who did not decide on the answer shows the following. The modified SERVPERF works better when the respondent has certain expectations about some aspect of the educational process. The modified SERVSAT works better when the respondent has a clear idea about real situation regarding some aspect of the educational process.
- 2023
Artificial Intelligence-Powered System for Identifying Bone Deterioration in Radiological Imaging(atveras jaunā logā)
Kaspars Sudars, Ivars Namatevs, Arturs Nikulins, Edgars Edelmers, Laura Neimane, Anda Slaidina, Oskars Radzins
International Workshop on Embedded Digital Intelligence (IWoEDI'2023)
Kopsavilkums
The aim of this research was to investigate the potential of deep convolutional neural networks (DCNN) for developing a reliable osteoporosis diagnostic tool using conebeam computed tomography (CBCT) scans of the mandible. The study utilized CBCT scans of patients' mandibular bone tissue and incorporated two pre-existing DCNN architectures derived from the ResNet-101 model. Findings from the study suggest that employing transfer learning methodologies can produce satisfactory outcomes in the creation of deep learning models for osteoporosis detection, even when the availability of mandibular CBCT image datasets is restricted.
- 2023
Modular Neural Networks for Osteoporosis Detection in Mandibular Cone-Beam Computed Tomography Scans(atveras jaunā logā)
Ivars Namatevs, Arturs Nikulins, Edgars Edelmers, Laura Neimane, Anda Slaidina, Oskars Radzins, Kaspars Sudars
Tomography, Sēj. 9 (5), lpp. 14, MDPI
Kopsavilkums
In this technical note, we examine the capabilities of deep convolutional neural networks (DCNNs) for diagnosing osteoporosis through cone-beam computed tomography (CBCT) scans of the mandible. The evaluation was conducted using 188 patients’ mandibular CBCT images utilizing DCNN models built on the ResNet-101 framework. We adopted a segmented three-phase method to assess osteoporosis. Stage 1 focused on mandibular bone slice identification, Stage 2 pinpointed the coordinates for mandibular bone cross-sectional views, and Stage 3 computed the mandibular bone’s thickness, highlighting osteoporotic variances. The procedure, built using ResNet-101 networks, showcased efficacy in osteoporosis detection using CBCT scans: Stage 1 achieved a remarkable 98.85% training accuracy, Stage 2 minimized L1 loss to a mere 1.02 pixels, and the last stage’s bone thickness computation algorithm reported a mean squared error of 0.8377. These findings underline the significant potential of AI in osteoporosis identification and its promise for enhanced medical care. The compartmentalized method endorses a sturdier DCNN training and heightened model transparency. Moreover, the outcomes illustrate the efficacy of a modular transfer learning method for osteoporosis detection, even when relying on limited mandibular CBCT datasets. The methodology given is accompanied by the source code available on GitLab. © 2023 by the authors.
Scopus procentile: 70
- 2023
Initial results of the development of intelligent non-invasive phenotyping of raspberries using machine learning and 3D imaging(atveras jaunā logā)
S. Strautiņa, I. Kalniņa, E. Kaufmane, K. Sudars, I. Namatēvs, J. Judvaitis, R. Balašs, A. Ņikuļins
Acta Horticulturae, Sēj. 4, lpp. 14, MDPI
Kopsavilkums
To distinguish candidate cultivars in fruit breeding, the characteristics of several thousand seedlings must be described and evaluated, which is done visually. Fruit counting and measuring in the field is currently the most used method. However, it is tedious and time-consuming and requires abundant manpower. In addition, visual evaluation is relatively subjective, and results may vary from different evaluators. The solution is to develop a machine vision-based method to automate and to intensify the fruit breeding process. In Latvia, the fruit breeding programme focuses on four fruit crops, including raspberries. Inclusion of raspberries (Rubus idaeus) in the breeding programme is related to the fact that there are no cultivars suitable for Latvian agroclimatic conditions. Breeding experience shows that the selections have insufficient ecological phenotype plasticity, which is a risk for cultivation of these cultivars in Latvian climate. In raspberries, yield depends on the sum of yield components, and each component is evaluated separately. The most important yield components are: number of laterals per shoot, number of flowers/fruits per lateral shoot, and average fruit mass. With the evaluation of these traits, image-based approaches to raspberry phenotyping are gaining momentum and provide fertile ground for non-invasive raspberry detection and categorization. The objective of this research is to develop a methodology and tools for non-invasive phenotyping of raspberry yield components based on red, green, blue (RGB) image colour value and 3D images, as well as provide descriptive and inferential statistics of raspberry cultivars. We propose a manually annotated 2D raspberry data set with ground truth region of interest (ROI)classifying labelled into five classes: “Buds”, “Flowers”, “Unripe Berries”, “Ripe Berries”, and “Damaged Berries” to verify and evaluate the raspberry detection problem using real-time deep neural network Yolo5. We also present three algorithms for 3D computational image processing to create 3D bounding boxes for raspberry parametrizing.
Indeksēts Scopus / WoS · Scopus procentile: 20