Computer Science Journal (AGH University of Science and Technology, Krakow)
Not a member yet
476 research outputs found
Sort by
A novel hybrid deep learning approach for 3d object detection and tracking in autonomous driving
Full text linkRecently Object detection and tracking using fusion of LiDAR and RGB camera for the autonomous vehicle environment is a challenging task. The existingworks initiates several object detection and tracking frameworks using ArtificialIntelligence (AI) algorithms. However, they were limited with high false positives and computation time issues thus lacking the performance of autonomousdriving environment. The existing issues are resolved by proposing HybridDeep Learning based Multi Object Detection and Tracking (HDL-MODT) using sensor fusion methods. The proposed work performs fusion of solid stateLiDAR, Pseudo LiDAR, and RGB camera for improving detection and trackingquality. At first, the multi-stage preprocessing is done in which noise removal isperformed using Adaptive Fuzzy Filter (A-Fuzzy). The pre-processed fused image is then provided for instance segmentation to reduce the classification andtracking complexity. For that, the proposed work adopts Lightweight GeneralAdversarial Networks (LGAN). The segmented image is provided for objectdetection and tracking using HDL. For reducing the complexity, the proposedwork utilized VGG-16 for feature extraction which forms the feature vectors.The features vectors are then provided for object detection using YOLOv4.Finally, the detected objects were tracked using Improved Unscented KalmanFilter (IUKF) and mapping the vehicles using time based mapping by considering their RFID, velocity, location, dimension and unique ID. The simulation ofthe proposed work is carried out using MATLAB R2020a simulation tool andperformance of the proposed work is compared with several metrics that showthat the proposed work outperforms than the existing works
Biometrics-Based Generation of Diffie-Hellman Key Exchange Parameters
Full text linkWhen two parties need to securely communicate over an insecure channel, Diffie-Hellman is often employed as the key exchange algorithm. This paper presents two novel approaches to generating Diffie-Hellman parameters for key exchange based on user biometrics, namely their fingerprint data. Fingerprint templates are extracted as bit strings via a fingerprint scanner and later used as inputs. In one approach, the whole fingerprint template is utilized as a user’s private key. In the second approach, fingerprint data is scrambled into smaller chunks and rearranged into two strings that serve as the user’s private key and the basis for prime p. Both approaches were implemented and tested experimentally. After analysis, the second approach that uses scrambled fingerprint data shows better execution times and improved security and usability considerations
The Ant Colony Optimization Algorithm Applied in Transport Logistics
Full text linkThe Vehicle Routing Problem belongs to graph optimization and its goal is to find shortest routes visiting a given set of customers with additional constraints present. The article presents the ant colony optimization metaheuristic which solves vehicle routing problems and its real-life application in transport logistics (finding routes for delivery companies). The metaheuristic generated high-quality solutions (superior to compared methods). Our tool is flexible and enables us to solve various variants of routing problems so it is well suited to specific needs of transportation companies
Using Deep Neural Networks to Improve the Precision of Fast-Sampled Particle Timing Detectors
Full text linkMeasurements from particle timing detectors are often affected by the time walk effect caused by statistical fluctuations in the charge deposited by passing particles. The constant fraction discriminator (CFD) algorithm is frequently used to mitigate this effect both in test setups and in running experiments, such as the CMS-PPS system at the CERN’s LHC. The CFD is simple and effective but does not leverage all voltage samples in a time series. Its performance could be enhanced with deep neural networks, which are commonly used for time series analysis, including computing the particle arrival time. We evaluated various neural network architectures using data acquired at the test beam facility in the DESY-II synchrotron, where a precise MCP (MicroChannel Plate) detector was installed in addition to PPS diamond timing detectors. MCP measurements were used as a reference to train the networks and compare the results with the standard CFD method. Ultimately, we improved the timing precision by 8% to 23%, depending on the detector\u27s readout channel. The best results were obtained using a UNet-based model, which outperformed classical convolutional networks and the multilayer perceptron
Enhanced bonobo optimizer for optimizing dynamic photovoltaic models
Full text linkBonobo optimizer (BO) is a novel metaheuristic algorithm motivated bythe social behaviour of the bonobos. This paper presents a quantum behaved bonobo optimization algorithm (QBOA) employing an innovative metaheuristic based on the reproductive strategies and social behavior of bonobos.Whereby, the quantum mechanics are embedded into the bonobo optimizerto direct the search agents through the search space. Accordingly, under thisquantum-behaved movement, the proposed QBOA’s exploitation capability ispromoted. The performance of the proposed QBOA is exhibited on CEC2005and CEC2019 benchmarks. Moreover, the QBOA algorithm was adapted tooptimize the dynamic photovoltaic models parameters. QBOA exhibits theefficiency and adequacy to solve various optimization problems based on experimental and comparison findings, as well as its ability to implement competitiveand promising results optimizing dynamic photovoltaic model
Generalizing Clustering Inferences with ML Augmentation of Ordinal Survey Data
Full text linkIn this paper, we attempt to generalize the ability to achieve quality inferences of survey data for a larger population through data augmentation and unification. Data augmentation techniques have proven effective in enhancing models\u27 performance by expanding the dataset\u27s size. We employ ML data augmentation, unification, and clustering techniques. First, we augment the \textit{limited} survey data size using data augmentation technique(s). Next, we carry out data unification, followed by clustering for inferencing. We took two benchmark survey datasets to demonstrate the effectiveness of augmentation and unification. One is on features of students to be entrepreneurs, and the second is breast cancer survey data. We compare the results of the inference obtained from the raw survey data and the newly converted data. The results of this study indicate that the machine learning approach, data augmentation with the unification of data followed by clustering, can be beneficial for generalizing the inferences drawn from the survey data
Machine Learning based Event Reconstruction for the MUonE Experiment
Full text linkAs currently operating high energy physics experiments produce a huge amount of data, new methods of fast and efficient event reconstruction are necessary to handle the immense load. Storing the unprocessed data is not feasible, forcing experiments to process the data online employing the algorithms of quality provided for the offline analysis, but within strict time constraints. In the MUonE experiment the machine learning based event reconstruction techniques are being implemented and tested in order to provide efficient online reduction of data and to maximize the statistical power of the final physics measurement
Clustering for Clarity: Improving Word Sense Disambiguation through Multilevel Analysis
Full text linkIn natural language processing, a critical activity known as word sense disambiguation (WSD) seeks to ascertain the precise meaning of an ambiguous wordin context. Traditional methods for WSD frequently involve supervised learning methods and lexical databases like WordNet. However, these methods fallshort in managing word meaning complexity and capturing fine-grained differences. In this paper, for increasing the precision and granularity of word sensedisambiguation we proposed multilevel clustering method that goes deeper in the nested levels as locate groups of linked context words and categorize themaccording to their word meanings. With this method, we can more effectively manage polysemy and homonymy as well as detect minute differences in meaning. An actual investigation of the SemCor corpus demonstrates the performance score of multilevel clustering in WSD. This proposed method successfullyseparated clusters and groups context terms according to how semantically related they are, producing improved disambiguation outcomes. A more detailedknowledge of word senses and their relationships may be obtained thanks to the clustering process, which makes it possible to identify smaller clusters inside larger clusters. The outcomes demonstrate how multilevel clustering may enhance the granularity and accuracy of WSD. Our solution overcomes the drawbacks of conventional approaches and provides a more fine-grained representation of word senses by combining clustering algorithms
Detection of Credit Card Fraud with Optimized Deep Neural Network in Balanced Data Condition
Full text linkDue to the huge number of financial transactions, it is almost impossible for humans to manually detect fraudulent transactions. In previous work, the datasets are not balanced and the models suffer from overfitting problems. In this paper, we tried to overcome the problems by tuning hyperparameters and balancing the dataset by hybrid approach using under-sampling and over-sampling techniques. In this study, we have observed that these modifications are effective to get better performance in comparison to the existing models. The MCC score is considered an important parameter in binary classification since it ensures the correct prediction of the majority of positive data instances and negative data instances. So, we emphasize on MCC score and our method achieved MCC score of 97.09%, which is far more (16 % approx.) than other state of art methods. In terms of other performance metrics, the result of our proposed model is also improved significantly
The Most Current Solutions using Virtual-Reality-Based Methods in Cardiac Surgery - A Survey
Full text linkThere is a widespread belief that VR technologies can provide controlled, multi-sensory, interactive 3D stimulus environments that engage patients in interventions and measure, record and motivate required human performance. In order to investigate state-of-the-art and associated occupations we provided a careful review of 6 leading medical and technical bibliometric databases. Despite the apparent popularity of the topic of VR use in cardiac surgery, only 47 articles published between 2002 and 2022 met the inclusion criteria. Based on them VR-based solutions in cardiac surgery are useful both for medical specialists and for the patients themselves. The new lifestyle required from cardiac surgery patients is easier to implement thanks to VR-based educational and motivational tools. However, it is necessary to develop the above-mentioned tools and compare their effectiveness with Augmented Reality (AR). With the aforementioned reasons, interdisciplinary collaboration between scientists, clinicians and engineers is necessary