Pētniecība
Publikācijas
729 publikācijas
- 2023
Towards Explainability of the Latent Space by Disentangled Representation Learning(atveras jaunā logā)
Ivars Namatēvs, Kaspars Sudars, Artūrs Ņikuļins, Anda Slaidiņa, Laura Neimane, Oskars Radziņš
Information Technology and Management Science, Sēj. 26 (1), lpp. 7, RTU press
Kopsavilkums
Deep neural networks are widely used in computer vision for image classification, segmentation and generation. They are also often criticised as “black boxes” because their decisionmaking process is often not interpretable by humans. However, learning explainable representations that explicitly disentangle the underlying mechanisms that structure observational data is still a challenge. To further explore the latent space and achieve generic processing, we propose a pipeline for discovering the explainable directions in the latent space of generative models. Since the latent space contains semantically meaningful directions and can be explained, we propose a pipeline to fully resolve the representation of the latent space. It consists of a Dirichlet encoder, conditional deterministic diffusion, a group-swap and a latent traversal module. We believe that this study provides an insight into the advancement of research explaining the disentanglement of neural networks in the community.
- 2023
Dataset on the Human Body as a Signal Propagation Medium for Body Coupled Communication(atveras jaunā logā)
Juris Ormanis, Vladislavs Medvedevs, Anastasija Sevcenko, Vladimir Aristov, Valters Abolins, Atis Elsts
Data in Brief, Elsevier
Kopsavilkums
Signal loss models are frequently utilized by wireless communication researchers and engineers to predict received signal strength, optimize system parameters, and conduct feasibility studies. However, novel communication methods such as Body-Coupled Communication (BCC) that are suitable for Body Area Networks formed by wearable devices currently lack readily available signal propagation models. In this data article, we present a galvanic-coupled BCC signal loss and bioimpedance dataset, which serves as a foundation for building such models. This extensive dataset consists of experimental data recorded from 30 volunteer test subjects. The experimental setup involves a tunable signal generator transmitting continuous wave signals, along with two oscilloscopes recording the transmitter-side and receiver-side voltages. From these measurements, we compute the signal loss over the body, and the transmitter-side impedance. The transmitted signal frequencies range from 50 kHz to 20 MHz, with discrete steps. The primary application of this dataset is to enable empirical-data-supported modeling in the human body as a BCC signal propagation medium, which will help to explore how the properties of the human body, the measurement locations, and the signal frequency impact the signal loss.
Scopus procentile: 72
- 2023
Yeast cell detection using fuzzy automatic contrast enhancement (FACE) and you only look once (YOLO)(atveras jaunā logā)
Zheng-Jie Huang, Brijesh Patel, Wei-Hao Lu, Tz-Yu Yang, Wei-Cheng Tung, Vytautas Bučinskas, Modris Greitans, Yu-Wei Wu & Po Ting Lin
Sci. Rep., Sēj. 13 (1), lpp. 16222, Nature Research
Kopsavilkums
In contemporary biomedical research, the accurate automatic detection of cells within intricate microscopic imagery stands as a cornerstone for scientific advancement. Leveraging state-of-the-art deep learning techniques, this study introduces a novel amalgamation of Fuzzy Automatic Contrast Enhancement (FACE) and the You Only Look Once (YOLO) framework to address this critical challenge of automatic cell detection. Yeast cells, representing a vital component of the fungi family, hold profound significance in elucidating the intricacies of eukaryotic cells and human biology. The proposed methodology introduces a paradigm shift in cell detection by optimizing image contrast through optimal fuzzy clustering within the FACE approach. This advancement mitigates the shortcomings of conventional contrast enhancement techniques, minimizing artifacts and suboptimal outcomes. Further enhancing contrast, a universal contrast enhancement variable is ingeniously introduced, enriching image clarity with automatic precision. Experimental validation encompasses a diverse range of yeast cell images subjected to rigorous quantitative assessment via Root-Mean-Square Contrast and Root-Mean-Square Deviation (RMSD). Comparative analyses against conventional enhancement methods showcase the superior performance of the FACE-enhanced images. Notably, the integration of the innovative You Only Look Once (YOLOv5) facilitates automatic cell detection within a finely partitioned grid system. This leads to the development of two models—one operating on pristine raw images, the other harnessing the enriched landscape of FACE-enhanced imagery. Strikingly, the FACE enhancement achieves exceptional accuracy in automatic yeast cell detection by YOLOv5 across both raw and enhanced images. Comprehensive performance evaluations encompassing tenfold accuracy assessments and confidence scoring substantiate the robustness of the FACE-YOLO model. Notably, the integration of FACE-enhanced images serves as a catalyst, significantly elevating the performance of YOLOv5 detection. Complementing these efforts, OpenCV lends computational acumen to delineate precise yeast cell contours and coordinates, augmenting the precision of cell detection.
Indeksēts Scopus / WoS · Scopus procentile: 100 · SNIP: 1.312
- 2023
Designing and Controlling a Self-Balancing Platform Mechanism Based on 3-RCC Spherical Parallel Manipulator(atveras jaunā logā)
Chen, Yen Jung, Wei-Cheng Tung, Wei-Rui Lee, Brijesh Patel, Vytautas Bučinskas, Modris Greitans, and Po Ting Lin
Robotic Systems and Applications, Sēj. 3 (1), lpp. 1-16, Extrica
Kopsavilkums
Motion control platforms have various applications in the manufacturing and automation industries. Different literature provides multiple issues related to the kinematics and dynamics of self-guided robots for transportation regarding platform balancing. Self-balancing platforms are utilized in many deliveries, stabilization, and transportation systems, and they are especially well suited for outdoor activities when the ground surface is not flat or structured. This paper describes developing a control technique for a self-balancing platform using the 3-RCC spherical parallel manipulator. This mechanism was designed to support an AGV (Automated Guided Vehicle) for transporting and lifting heavy weights for industrial applications. The AGV carries a robotic arm on top for different tasks. When the AGV encounters a steep slope or a rough surface, the AGV tilts, and the robotic arm’s performance is significantly affected. So, this study gives a solution to avoid these circumstances with a novel approach for the platform’s self-balancing mechanism consisting of a 3-RCC spherical parallel manipulator. Real-time stabilization and kinematics analysis methods are used to achieve the self-balancing system of the platform. When both methods are observed through different tilting angles for automation stability, Kinematic analysis performs more efficiently with less time duration when compared with the real-time stabilization method.
- 2023
Automatic Cell Recognition in Microscopic Images Based on Contrast Enhancement and YOLO
Wei-Cheng Tung, Brijesh Patel, Vytautas Bučinskas, Andrius Dzedzickis, Modris Greitans, Po Ting Lin
2023 International Conference on Advanced Robotics and Intelligent Systems (ARIS 2023)
Kopsavilkums
The recognition of living cells in microscopic images is an important part of creating new types of autonomous, flexible microrobots for the manipulation of individual living cells. However, such systems are not widespread since the cost of designing them is high and they require significant human resources. Manual recognition of yeast cells in microscopic images is time-consuming and subject to human error. In this paper, we proposed a system for automatic cell recognition in microscopic images of living cells. The system utilized the You Only Look Once Version 7 (YOLOv7) algorithm for segmenting and marking living cells in microscopic images. To improve the stability of model training, validation, and testing, we enhanced the contrast of the microscopic images. Our experimental results showed that the proposed system achieved high accuracy and robustness in recognizing yeast cells in microscopic images.
- 2023
Wireless induction charging station for industrial application(atveras jaunā logā)
Pēteris Aizpurietis, Didzis Lapsa, Niklāvs Barkovskis, Arnis Salmiņš, Rims Janeliukštis, Kaspars Ozols
2023 Workshop on Microwave Theory and Technology in Wireless Communications (MTTW), lpp. 62-66, IEEE
Kopsavilkums
Wiring for charging LiPo batteries poses problems due to electromagnetic interference and limitations of wire length. In this work, a method for charging wireless sensor network wirelessly using NRF52840 is developed for both transmitting and receiving side of equipment. A Qi-compliant WE760308103202 wireless charging coil receiver is used and its performance evaluated to estimate optimal parameters of wireless charging, such as load and resonance frequency detuning for receiver. By tuning the frequency of operation, it is possible to utilize generic electronic components with reduced costs.
Indeksēts Scopus / WoS
- 2023
Method for Muscle Fatigue Detection Using Inertial Sensors(atveras jaunā logā)
Banga, B., Katashev, A., Greitāns, M.
In Nordic-Baltic Conference on Biomedical Engineering and Medical Physics. 2023., Sēj. 89, lpp. 25-32, Springer
Kopsavilkums
Muscle fatigue is a common symptom that many people experience and is associated with difficulties in voluntary movement, which can lead to injuries. Currently, surface electromyography (sEMG) is considered the gold standard for muscle fatigue estimation, but its accuracy can be impacted by various factors. Therefore, new methods, such as the use of inertial sensors (IMU), are being introduced. This study aimed to explore the relationship between muscle fatigue and biomechanical parameters using inertial sensors and sEMG as a validation tool. Four participants performed an elbow flexion exercise, and the data from IMU sensor nodes and sEMG were collected. The results showed that there were correlations between the electrical activity of m. biceps brachii and rotation angles of the forearm and upper arm. Additionally, an increase in motion amplitude deviation was found to be a potential indicator of muscle fatigue. These findings suggest that inertial sensors can be used as an alternative to sEMG for detecting muscle fatigue, which has potential implications for injury prevention and rehabilitation. However, further research with a larger sample size is needed to validate these findings.
Indeksēts Scopus / WoS · Scopus procentile: 21
- 2023
Transmission Sensing of Dielectric Objects for Determining the Geometric Shape of the Cross Section and the Detection of Internal Inhomogeneities(atveras jaunā logā)
V. Aristov
Automatic Control and Computer Sciences, Sēj. 57, lpp. 196-202, Springer
Kopsavilkums
In this work, the possibility of determining the geometric shape of the cross section of dielectric objects in the case of a priori information on their internal homogeneity or detecting integral internal inhomogeneities in the case of a priori information on the geometric shape of the cross section via the transmission sensing of these objects with ultra-wideband pulse radars with a central frequency in the spectrum of pulses on the order of 1 GHz is investigated.
Indeksēts Scopus / WoS · Scopus procentile: 40
- 2023
Remote Complex Resistance Measurement(atveras jaunā logā)
V. Aristov
Automatic Control and Computer Sciences, Sēj. 57 (5), lpp. 469-472
Kopsavilkums
In the proposed work we investigate the possibility of measuring a complex resistance remote by several meters, in particular the parallel connection of an active resistance and a capacitor. The measurement is made at a frequency of one megahertz. Since the unknown complex resistance is connected to the measuring equipment by a radio-frequency cable, the theory of long lines is used in deriving the formula. Examples of practical application of the algorithm for measuring human skin resistance are given.
Indeksēts Scopus / WoS · Scopus procentile: 40
- 2023
An experimental study: RF Fingerprinting of Bluetooth devices(atveras jaunā logā)
Artis Rušiņš, Krišjānis Nesenbergs, Deniss Tiščenko, Pēteris Paikens
2023 International Conference on Embedded Wireless Systems and Networks (EWSN '23), ACM
Kopsavilkums
This paper presents an experimental study on radio frequency (RF) fingerprinting of Bluetooth Classic devices. Our research aims to provide a practical evaluation of the possibilities for RF fingerprinting of everyday Bluetooth connected devices that may cause privacy risks. We have built an experimental setup for recording Bluetooth connection in a radio frequency isolated environment using commercially available SDR (software defined radio) systems, extracted fingerprints of the Bluetooth radio data in the form of carrier frequency offset and scaling factor from 6 different devices, and performed k-nearest neighbors (kNN) classification achieving 84 % accuracy. The experiment demonstrates that no matter what privacy measures are being taken in the protocol layer, the physical layer leaks significant information about the device to unauthorized listeners. In the context of the ever-growing Bluetooth device market, this research serves as a clarion call for device manufacturers, regulators, and end-users to acknowledge the privacy risks posed by RF fingerprinting and lays a foundation for more sizeable Bluetooth fingerprinting analysis research.
- 2023
A study on Automated railway level crossing control system using FMCW radar for accident prevention(atveras jaunā logā)
Kristaps Greitans
2023 International Conference on Applied Electronics (AE), Pilsen, Czech Republic, 2023
Kopsavilkums
This study paper explores the implementation of Intelligent Transportation Systems (ITS) in railway transportation, with a focus on level crossings, which have been a significant issue in railway safety. The paper discusses the limitations of current technologies such as laser technology and video surveillance and presents a novel solution utilizing a 79GHz Frequency Modulated Continuous Wave (FMCW) radar in combination with a CCTV camera for a dual sensor module. The FMCW radar is shown to be more accurate and cost-effective than previous technologies and can detect both large and small objects in the defined area at level crossings and difficult weather situations. The paper also presents a system architecture that takes into account vehicle detection time, which is crucial in reducing the risk of accidents and improving the safety of level crossings. This study aimed to compare the performance of different object detection modules under varying weather conditions. Three object detection methods were tested: a camera with the object detection algorithm Yolov5, FMCW radar with a built-in tracker, and manual observation. Four measurement sessions were conducted, each with a different combination of weather and time of day. The results showed that all modules performed similarly well under clear weather conditions, with the camera and radar modules detecting all vehicles approaching. However, under rainy and dark conditions, the radar module outperformed the camera, detecting 5 more objects.
Indeksēts Scopus / WoS
- 2023
BEAM: Body Coupled Communication Enabled Amplitude Modulation for Skinput application
Juris Ormanis, Vladislavs Medvedevs, Jānis Judvaitis
International Conference on Embedded Wireless Systems and Networks (EWSN '23)
Kopsavilkums
Body Coupled Communication (BCC) technology has the potential to revolutionise the wearable device field by provid- ing substantial advantages in security, privacy and usability, yet understanding of the factors that influence signal path loss in BCC is crucial for developing accurate measurement techniques and refining the communication system. In this paper a new approach in proposed taking advantage of vary- ing signal loss on touch to enable user input detection and interpretation, it demonstrates high accuracy and reliability. This technology can potentially provide new opportunities for visually impaired or disabled individuals by designating specific body locations as scenario activation triggers and en- abling tactile and intuitive interaction with the technological environment. To study the impact of human touch on signal path loss multiple experiments were carried out and the find- ings enhance the understanding of the factors affecting BCC performance and provides an insight for further research.
Indeksēts Scopus / WoS · Scopus procentile: 35
- 2023
Poster: IoT-Edge-Cloud Continuum data flow validation tool
Jānis Judvaitis, Krišjānis Nesenbergs
Proceedings of Embedded Wireless Systems and Networks 2023, Association for Computing Machinery
Kopsavilkums
IoT-Edge-Cloud continuum is becoming a de facto for any IoT or AI-related service based on how the gathered data are processed and stored, with the new emergence of Edge computing. Since all of the citizens should take advantage of public services created in the IoT Edge-Cloud continuum the trust, security, and privacy questions are of utmost importance not only for the developers but also for other involved actors, such as owners, operators, and users. In this poster we propose a novel IoT-Edge-Cloud Continuum data flow validation tool aiming to assist the involved parties in aiming for the best possible outcome for society as a whole, contributing to the development of the IoT-Edge-Cloud continuum services as well as providing a mechanism for validation of data management mechanisms in theory as well as in practice.
Indeksēts Scopus / WoS
- 2023
Numerical Modelling of a Turbine Flow Meter Used as Part of the Hydrogen Compressor System
Orlova S., Mezeckis N., Tejesh N.D., Vasudev V.P.K., Upnere S
Latv. J. Phys. Tech. Sci.
- 2023
Compression of hydrogen gas for energy storage: a review
Orlova S., Mezeckis N., Vasudev V.P.K.
Latv. J. Phys. Tech. Sci.
- 2023
Implementing European Union Provisions and Enabling Frameworks for Renewable Energy Communities in Nine Countries: Progress, Delays, and Gaps(atveras jaunā logā)
Krug M. et.al. (15 authors including Klavs G.)
Sustainability, Sēj. 15 (11), lpp. 8861
- 2023
Wetland Change Detection Using Sentinel-2 in the Part of Latvia(atveras jaunā logā)
Andris Skromulis, Juris Breidaks, Mārtiņš Puķītis
Environment. Technology. Resources. Proceedings of the International Scientific and Practical Conference June 15-16, 2023, Sēj. 1, Rezekne Academy of Technologies
Kopsavilkums
In the article, the possible impact of changes on wetland were analysed by the semi-supervised classification method of statistical analysis. The Sentinel-2 raw data between two different seasons are combined together. The data preparation is shortly described in the article. Data is clustered with unsupervised method. The article describes a supervised method – how data credibility and classification can be estimated if its reference is poor quality.
Indeksēts Scopus / WoS
- 2023
Classification of satellite images using Dynland technology(atveras jaunā logā)
Martins Pukitis, Ints Mednieks
Ninth International Conference on Remote Sensing and Geoinformation of the Environment (RSCy2023), Sēj. 127860, Proc. SPIE 12786
Kopsavilkums
Satellite images are widely used for land cover classification into predefined categories. To perform it in a practicable way, field data are needed to train and test a classifier. The collection of accurate field data requires significant resources, and image data of a category may have sophisticated distribution in the feature space. “Dynland” technology was developed to overcome these problems and offer a robust and practical way to classify such images. It is based on a nonparametric clustering method applicable to data with different distributions with a following assignment of classes to clusters based on their overlapping with available reference data which can be scarce and imprecise. Dynland clustering algorithm provides meaningful clusters but requires huge computational resources currently limiting the size of processed images. To overcome this problem, we propose to apply a sampling procedure i.e. use each 𝑛-th pixel on both axes for clustering and distribute the remaining pixels to created clusters based on nearest neighbors’ search. However, such an approach reduces the classification accuracy; the analysis results of this reduction will be presented. Another way how to deal with images of large areas is to split them into fragments. To classify a fragment, reference data for all categories of interest should be available for the area it represents. This limitation should be resolved so that the available reference data are used for areas not limited to one fragment. To deal with that, a method is proposed allowing to exploit the classification result of a neighboring fragment for which the reference data were available in full. The proposed solutions widened the applicability of the Dynland technology e.g. enabled the production of a classified map of mires and peatlands for the whole territory of Latvia.
Indeksēts Scopus / WoS
- 2022
Improving Performance of the PRYSTINE Traffic Sign Classification by Using a Perturbation-Based Explainability Approach(atveras jaunā logā)
Sudars, Kaspars, Ivars Namatēvs, and Kaspars Ozols
Journal of Imaging, Sēj. 8 (2), lpp. 30, 1.339
Kopsavilkums
Model understanding is critical in many domains, particularly those involved in high-stakes decisions, e.g., medicine, criminal justice, and autonomous driving. Explainable AI (XAI) methods are essential for working with black-box models such as convolutional neural networks. This paper evaluates the traffic sign classifier of the Deep Neural Network (DNN) from the Programmable Systems for Intelligence in Automobiles (PRYSTINE) project for explainability. The results of explanations were further used for the CNN PRYSTINE classifier vague kernels’ compression. Then, the precision of the classifier was evaluated in different pruning scenarios. The proposed classifier performance methodology was realised by creating an original traffic sign and traffic light classification and explanation code. First, the status of the kernels of the network was evaluated for explainability. For this task, the post-hoc, local, meaningful perturbation-based forward explainable method was integrated into the model to evaluate each kernel status of the network. This method enabled distinguishing high- and low-impact kernels in the CNN. Second, the vague kernels of the classifier of the last layer before the fully connected layer were excluded by withdrawing them from the network. Third, the network’s precision was evaluated in different kernel compression levels. It is shown that by using the XAI approach for network kernel compression, the pruning of 5% of kernels leads to a 2% loss in traffic sign and traffic light classification precision. The proposed methodology is crucial where execution time and processing capacity prevail.
Indeksēts Scopus / WoS · Q1/Q2 žurnāls · Scopus procentile: 82 · SNIP: 1.339
- 2022
Smart Industrial Robot Control Trends, Challenges and Opportunities within Manufacturing(atveras jaunā logā)
Janis Arents, Modris Greitans
Applied Sciences, Sēj. 12 (2), lpp. 937, MDPI
Kopsavilkums
Industrial robots and associated control methods are continuously developing. With the recent progress in the field of artificial intelligence, new perspectives in industrial robot control strategies have emerged, and prospects towards cognitive robots have arisen. AI-based robotic systems are strongly becoming one of the main areas of focus, as flexibility and deep understanding of complex manufacturing processes are becoming the key advantage to raise competitiveness. This review first expresses the significance of smart industrial robot control in manufacturing towards future factories by listing the needs, requirements and introducing the envisioned concept of smart industrial robots. Secondly, the current trends that are based on different learning strategies and methods are explored. Current computer-vision, deep reinforcement learning and imitation learning based robot control approaches and possible applications in manufacturing are investigated. Gaps, challenges, limitations and open issues are identified along the way.
Indeksēts Scopus / WoS · Q1/Q2 žurnāls · Scopus procentile: 74 · SNIP: 1.068