Conference Papers (KIV)
Permanent URI for this collection
Browse
Recent Submissions
Item TVMC: Time-Varying Mesh Compression Using Volume-Tracked Reference Meshes(Association for Computing Machinery, 2025) Chen, Guodong; Hácha, Filip; Váša, Libor; Dasari, MalleshamTime-varying meshes (TVMs), characterized by their varying connectivity and number of vertices, hold significant potential in AR/VR applications. However, their practical use is challenging due to their large file sizes and the complexity of time-varying topology. Many time-varying mesh compression methods attempted to exploit redundancy between consecutive meshes to compress TVMs more efficiently, however, most face difficulties in establishing stable vertex and surface correspondence between the frames of a TVM. We propose TVMC, a novel TVM compression method that leverages volume tracking and extracts high-quality reference meshes for inter-frame prediction. Specifically, we use as-rigid-as-possible volume tracking to align consecutive TVMs and track volume centers, followed by multidimensional scaling to refine reference centers. This allows us to precisely deform a group of frames to the reference space and extract the reference mesh which is then deformed to approximate each mesh in the group to get displacement fields for TVM compression. Extensive experiments show that TVMC outperforms state-of-the-art methods (e.g., Google Draco, V-DMC 4.0, etc.), with bitrates of 4-6 Mbps compared to 9-12 Mbps for Draco and 10-15 Mbps for V-DMC 4.0. It reduces the decoding time by 66.1% compared to Draco and enables an increased group of frames (up to 15) without significant distortion.
Item Large Language Models for Summarizing Czech Historical Documents and Beyond(ScitePress, 2025) Tran, Václav; Šmíd, Jakub; Martínek, Jiří; Lenc, Ladislav; Král, PavelText summarization is the task of shortening a larger body of text into a concise version while retaining its essential meaning and key information. While summarization has been significantly explored in English and other high-resource languages, Czech text summarization, particularly for historical documents, remains underexplored due to linguistic complexities and a scarcity of annotated datasets. Large language models such as Mistral and mT5 have demonstrated excellent results on many natural language processing tasks and languages. Therefore, we employ these models for Czech summarization, resulting in two key contributions: (1) achieving new state-of-the-art results on the modern Czech summarization dataset SumeCzech using these advanced models, and (2) introducing a novel dataset called Posel od Čerchova for summarization of historical Czech documents with baseline results. Together, these contributions provide a great potential for advancing Czech text summarization and open new avenues for research in Czech historical text processing.Item Advancing Cross-Lingual Aspect-Based Sentiment Analysis with LLMs and Constrained Decoding for Sequence-to-Sequence Models(ScitePress, 2025) Šmíd, Jakub; Přibáň, Pavel; Král, PavelAspect-based sentiment analysis (ABSA) has made significant strides, yet challenges remain for low-resource languages due to the predominant focus on English. Current cross-lingual ABSA studies often centre on simpler tasks and rely heavily on external translation tools. In this paper, we present a novel sequence-to sequence method for compound ABSA tasks that eliminates the need for such tools. Our approach, which uses constrained decoding, improves cross-lingual ABSA performance by up to 10%. This method broadens the scope of cross-lingual ABSA, enabling it to handle more complex tasks and providing a practical, efficient alternative to translation-dependent techniques. Furthermore, we compare our approach with large language models (LLMs) and show that while fine-tuned multilingual LLMs can achieve comparable results, English centric LLMs struggle with these tasks.Item LACA: Improving Cross-lingual Aspect-Based Sentiment Analysis with LLM Data Augmentation(Association for Computational Linguistics, 2025) Šmíd, Jakub; Přibáň, Pavel; Král, PavelCross-lingual aspect-based sentiment analysis (ABSA) involves detailed sentiment analysis in a target language by transferring knowledge from a source language with available annotated data. Most existing methods depend heavily on often unreliable translation tools to bridge the language gap. In this paper, we propose a new approach that leverages a large language model (LLM) to generate high-quality pseudo-labelled data in the target language without the need for translation tools. First, the framework trains an ABSA model to obtain predictions for unlabelled target language data. Next, LLM is prompted to generate natural sentences that better represent these noisy predictions than the original text. The ABSA model is then further fine-tuned on the resulting pseudo-labelled dataset. We demonstrate the effectiveness of this method across six languages and five backbone models, surpassing previous state-of-the-art translation-based approaches. The proposed framework also supports generative models, and we show that fine-tuned LLMs outperform smaller multilingual models.Item Fast Approximate Symmetry Plane Computation as a Density Peak of Candidates(SciTePress, 2025) König, Alex; Váša, LiborSymmetry is a common characteristic exhibited by both natural and man-made objects. This property can be used in various applications in computer vision and computer graphics. There are various types of symmetries, amongst the most prominent belong reflection symmetries and rotation symmetries. In this paper, a method focusing on the fast detection of approximate reflection symmetry of a 3D point cloud with respect to a plane is proposed. The method is based on the creation of a set of candidates that are represented as rigid transformations, and have assigned weights, reflecting the estimated quality of the candidate. The final symmetry plane corresponds to a density peak in the transformation space. The method is demonstrated to be able to find symmetry planes in various objects in 3D, with its main benefit being the speed of the computation.Item Online Simulation of a Diabetic-Patient Metabolism and a Realistic Meal Composition(IEEE, 2025) Koutný, Tomáš; Úbl, MartinThere are in-silico models to aid both research and treatment of patients with diabetes. In general, these models consider insulin and carbohydrate content of the ingested food only. By considering complete meal composition, i.e., carbohydrates, protein, fat and fiber, we can substantially improve the metabolic simulation to obtain considerably more precise metabolic responses to meal, drugs and physical activity. To do so, we integrated the SmartCGMS framework for glucose level monitoring and control with the state-of-the-art Sirael virtual metabolic machine, which is capable of simulatingmetabolic responses to a realistic meal composition. Specifically, we adapted the capability of building SmartCGMS for low- power devices to produce a WebAssembly instead of targeting alow-power device. At diabetes.zcu.cz, there is a live demonstration of our work, which can execute online, in a web browser. It demonstrates, how a web browser brings the simulation to patients for education, to improve their outcomes by increasing their motivation via learning. In the wider context, he presented work extends research options for both biomedical engineers, physicians and health care professionals, which educate patients.Item Symmetry Detection and Symmetrization in Cellular Automata(SciTePress, 2025) Gregor, Vít; Kolingerová, IvanaSymmetry is the important property of many geometric objects. Our work analyzes the symmetry in two-dimensional objects created using the cellular automata in the context of the initial configurations and rules of the automata. Symmetry of basic geometric shapes, such as circles, rectangles, and curves, mapped into the cells of the automaton, is analyzed in this paper. Also, the symmetry of random objects is analyzed. This paper also describes the method for centroidal symmetry detection and axial symmetry detection in the cellular automata, and it also brings the approach of using the cellular automata for object symmetrization by comparing it with objects in the library of symmetrical objects.Item Finding the Shear Reflection Symmetry Plane in a 3D Point Cloud(ScitePress, 2025) Poór, Vítek; Kolingerová, Ivana; Strnad, DamjanMany objects, namely man-made ones, show signs of various types of symmetry. The most common type perceived by humans is reflection symmetry to some plane. When detecting the symmetry for geometric models, the existing algorithms look for orthogonal reflection symmetry. However, the models can be sheared, therefore, algorithms detecting shear reflection symmetry would be useful. In this paper, we propose an algorithm for detecting the plane of shear reflection symmetry in a 3D point cloud on condition that the shear was done in one of the coordinate axes.Item A Library for General Genetic Algorithm with Sequential/Parallel Fitness Function Calculation(IEEE, 2025) Potužák, Tomáš; Křenek, FilipIIn this paper, we describe a library for Java, which enables easy usage of a genetic algorithm in a project. The library – Genetic Algorithm Library for Java (GAL4J) – contains multiple utilizable variants of individual parts of the genetic algorithm and also enables parallel computation of the fitness values of the individuals. The usability of the library was demonstrated on two classical problems. The speed of the GAL4J genetic algorithm using sequential and parallel fitness function calculation was investigated. The GAL4J was also compared to two existing GA libraries of third parties.
Item Augmentation of Motor Imagery Data for Brain-Controlled Robot-Assisted Rehabilitation(SCITEPRESS – Science and Technology Publications, Lda., 2024) Mouček, Roman; Kodera, Jakub; Mautner, Pavel; Průcha, JaroslavBrain-controlled robot-assisted rehabilitation is a promising approach in healthcare that can potentially and in parallel improve and partly automate the rehabilitation of motor apparatus and related brain structures responsible for movement. However, building a real-world rehabilitation system has many challenges and limitations. One of these challenges is the small size of the data that can be collected from the target group of people recovering from injured motor functions to train deep learning models recognizing motor imagery patterns. Therefore, the primary experiments with data augmentation and classification results over the collected and augmented dataset are presented.Item Sleep Apnea Detection from Single-Lead ECG Signal Using Hybrid Deep CNN(Springer Nature Singapore, 2025) Pham, Duc Thien; Mouček, RomanSleep apnea (SA) is a prevalent disorder that disrupts breathing during sleep, posing risks to multiple organs and potentially causing sudden death. The electrocardiogram (ECG) is vital for diagnosing SA due to its ability to identify irregular heart activity. This study introduces hybrid CNN models designed to automatically detect SA using a single-lead ECG signal. We validated our method through experiments with the Physionet Apnea-ECG dataset, which contains 70 single-lead ECG recordings annotated by medical professionals. Our results surpass the current state-of-the-art methods in accurately detecting SA from single-lead ECG signals, achieving an accuracy of 91.4% for per-segment classification and 100% for per-recording classification.Item Improvement of Replicas Placement in DFS(Springer, 2025) Pešička, Ladislav; Matějka, LubošDistributed File Systems (DFS) have been used to store and access data for many years. Recently, with the increase in the number of mobile devices and their connection speeds using 5G, there is a need to optimize the structure of DFS to meet these demands. By placing data replicas appropriately, the performance of the DFS system can be improved. This paper discusses the effect of replica placement on improving DFS parameters.Item Projective Geometric Algebra - Barycentric and Plücker Coordinates Computation(University of Defence, 2022) Skala, VáclavThis paper presents the computation of the barycentric coordinates and Plücker coordinates using the projective extension of the Euclidean space and geometric algebra. Using the projective extension, it also presents a relationship between linear systems of equations Ax=b and Ax=0 using the projective extension. An application of the principle of duality enables solving dual problems efficiently. The given approach uses vector notation leading to efficient implementation on GPU or efficient use of SSE instructions. As the presented approach is based on projective notation, the division operation is postponed and the proposed method leads to higher computational robustnessItem Current Trends in Automated Test Case Generation(PTI, 2023) Potužák, Tomáš; Lipka, RichardThis paper is a survey of existing literature of the last two decades that deals with test data generation or with tests based on it. This survey is not a systematic literature review and it does not try to answer specific scientific questions formulated in advance. Its purpose is to map and categorize the existing methods and to summarize their common features. Such a survey can be helpful for any teams developing their methods for test data generation as it can be a starting point for the exploration of related work.Item Classification of Event-Related Potential Signals with a Variant of UNet Algorithm Using a Large P300 Dataset(Springer, 2023) Khoshkhooy Titkanlou, Maryam; Mouček, RomanEvent-related potential signal classification is a really difficult challenge due to the low signal-to-noise ratio. Deep neural networks (DNN), which have been employed in different machine learning areas, are suitable for this type of classification. UNet (a convolutional neural network) is a classification algorithm proposed to improve the classification accuracy of P300 electroencephalogram (EEG) signals in a non-invasive brain-computer interface. The proposed UNet classification accuracy and precision were 64.5% for single-trial classification using a large P300 dataset of school-aged children, including 138 males and 112 females. We compare our results with the related literature and discuss limitations and future directions. Our proposed method performed better than traditional methods.Item Tunneling and Replication in Hierarchical DFS(Springer, 2023) Pešička, Ladislav; Matějka, LubošNowadays, distance learning, 5G mobile networks, and 4K cameras in mobile phones generate a large amount of data consumed on demand. The storage systems must fulfill these demands. This chapter presents a counting algorithm to optimize access in the hierarchical distributed file system.Different access strategies, such as tunneling, replication, and reconnection, are discussed and used to optimize file access.Item A generic model of hyperspace curvature preservation for a dynamic radial basis function implicit surface(University of West Bohemia, 2023) Červenka, MartinThe primary objective of the proposed methodology is to be implemented in the domain of computerized muscle modelling. However, preserving shape under different deformation scenarios is crucial across numerous research domains. Consequently, this paper briefly exposes the problem in a broader context. Multiple approaches exist to describe the surface, each with inherent strengths and limitations. In this context, the implicit surface utilized for deformation will be characterized using a collection of radial basis functions (RBFs). Although the choice of RBF may vary depending on the specific domain, this paper employs the widely recognized Gaussian RBF. This selection ensures a smooth and infinitely differentiable surface model, which is advantageous for the intended purpose.Item Genetic-Algorithm-Based Road Traffic Network Division with Improved Refining(IEEE, 2024) Potužák, TomášIn this paper, a method for road traffic network division for distributed and/or parallel road traffic simulation is described. Similarly to its former version, it employs multi-level graph partitioning scheme with graph coarsening, where a genetic algorithm is used for initial partitioning. The version of the method described in this paper (Improved Dividing Genetic Algorithm with Graph Coarsening and Improved Refining – IDGA-GC-IR) incorporates improved refining phase. In order to investigate its performance, it was compared to a third-party division method using testing on five road traffic networks. The tests showed that the IDGA-GC-IR method outperforms the compared third-party method in some important aspects of the division (e.g., number of divided traffic lanes), but is significantly slower.Item A New Highly Efficient Preprocessing Algorithm for Convex Hull, Maximum Distance and Minimal Bounding Circle in E2: Efficiency Analysis(Springer, 2024) Skala, VáclavThis contribution describes an efficient and simple preprocessing algorithm for finding a convex hull, maximum distance of points or convex hull diameter, and the smallest enclosing circle in E2. The proposed algorithm is convenient for large data sets with unknown intervals and ranges of the data sets. It is based on efficient preprocessing, which significantly reduces points used in final processing by standard available algorithms.Item Design of BCI-Based Exoskeleton System for Knee Rehabilitation(ScitePress, 2024) Khoshkhooy Titkanlou, Maryam; Pham, Duc Thien; Mouček, RomanInjuries of the lower limb, particularly the knee, usually require several months of rehabilitation. Exoskeletons are great tools supporting the rehabilitation process; their research and suitable practical use are at the center of interest of researchers and physiotherapists. This paper focuses on designing a brain-computer-interface (BCI)-controlled exoskeleton for knee rehabilitation. It includes reviewing and selecting electroencephalography (EEG) acquisition methods, BCI paradigms, current acquisition devices, signal classification methods and techniques, and the target group of people for whom the exoskeleton will be suitable. Finally, the preliminary proposal of the exoskeleton is provided.
