TY - THES AU - Rácz, Gergely Ferenc TI - Isotropic 3D sampling and signal reconstruction PB - Budapesti Műszaki és Gazdaságtudományi Egyetem PY - 2022 SP - 100 UR - https://m2.mtmt.hu/api/publication/34738092 ID - 34738092 LA - English DB - MTMT ER - TY - CHAP AU - Rácz, Gergely Ferenc AU - Kacsó, Ágota Enikő AU - Tóth, Márton József AU - Szirmay-Kalos, László ED - Szirmay-Kalos, László ED - Kiss, Bálint TI - Dynamic PET Reconstruction from Sinogram Data using Neural Networks T2 - Proceedings of the Workshop on the Advances of Information Technology 2020 : WAIT 2020 PB - Budapesti Műszaki és Gazdaságtudományi Egyetem, Villamosmérnöki és Informatikai Kar CY - Budapest SN - 9789634218029 PY - 2020 SP - 99 EP - 104 PG - 6 UR - https://m2.mtmt.hu/api/publication/31205325 ID - 31205325 LA - English DB - MTMT ER - TY - CHAP AU - Rácz, Gergely Ferenc AU - Kacsó, Ágota Enikő AU - Tóth, Márton József AU - Tóth, Balázs György ED - Kiss, Bálint ED - Szirmay-Kalos, László TI - Fast and accurate initial parameter fitting in Positron Emission Tomography using Neural Networks T2 - Proceedings of the Workshop on the Advances of Information Technology: WAIT 2019 PB - Budapesti Műszaki és Gazdaságtudományi Egyetem CY - Budapest SN - 9789633133101 PY - 2019 SP - 65 EP - 71 PG - 7 UR - https://m2.mtmt.hu/api/publication/30717085 ID - 30717085 LA - English DB - MTMT ER - TY - CHAP AU - Tóth, Balázs György AU - Tóth, Márton József AU - Kacsó, Ágota Enikő AU - Rácz, Gergely Ferenc AU - Szirmay-Kalos, László TI - Controlling TV Regularization with Deep Learning T2 - 2018 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2018 PB - Institute of Electrical and Electronics Engineers (IEEE) SN - 1538684942 PY - 2018 SP - 1 EP - 5 PG - 5 DO - 10.1109/NSSMIC.2018.8824545 UR - https://m2.mtmt.hu/api/publication/30371042 ID - 30371042 LA - English DB - MTMT ER - TY - CHAP AU - Rácz, Gergely Ferenc AU - Kacsó, Ágota Enikő AU - Tóth, Márton József AU - Tóth, Balázs György ED - Szirmay-Kalos, László ED - Renner, Gábor TI - Enhanced PET Reconstruction with Neural Networks T2 - IX. magyar számítógépes grafika és geometria konferencia, GRAFGEO 2018 PB - Neumann János Számítógép-tudományi Társaság CY - Budapest SN - 9789633132821 PY - 2018 SP - 125 EP - 131 PG - 7 UR - https://m2.mtmt.hu/api/publication/3381208 ID - 3381208 LA - English DB - MTMT ER - TY - JOUR AU - Rácz, Gergely Ferenc AU - Csébfalvi, Balázs TI - Cosine-Weighted B-Spline Interpolation on the Face-Centered Cubic Lattice JF - COMPUTER GRAPHICS FORUM J2 - COMPUT GRAPH FORUM VL - 37 PY - 2018 IS - 3 SP - 503 EP - 511 PG - 9 SN - 0167-7055 DO - 10.1111/cgf.13437 UR - https://m2.mtmt.hu/api/publication/3368783 ID - 3368783 AB - Cosine-Weighted B-spline (CWB) interpolation [Cse13] has been originally proposed for volumetric data sampled on the Body-Centered Cubic (BCC) lattice. The BCC lattice is well known to be optimal for sampling isotropically band-limited signals above the Nyquist limit. However, the Face-Centered Cubic (FCC) lattice has been recently proven to be optimal for low-rate sampling. The CWB interpolation is a state-of-the-art technique on the BCC lattice, which outperforms, for example, the previously proposed box-spline interpolation in terms of both efficiency and visual quality. In this paper, we show that CWB interpolation can be adapted to the FCC lattice as well, and results in similarly isotropic signal reconstructions as on the BCC lattice. LA - English DB - MTMT ER - TY - CHAP AU - Rácz, Gergely Ferenc AU - Kacsó, Ágota Enikő AU - Tóth, Márton József AU - Tóth, Balázs György ED - Kiss, Bálint ED - Szirmay-Kalos, László TI - PET Image Denoising using a Deep Neural Network T2 - Proceedings of the Workshop on the Advances of Information Technology PB - BME Irányítástechnika és Informatika Tanszék CY - Budapest SN - 9789633132784 PY - 2018 SP - 110 EP - 116 PG - 7 UR - https://m2.mtmt.hu/api/publication/3330835 ID - 3330835 AB - Neural networks are used in a wide range of image processing applications. They are able to solve complex tasks, e.g., image classification, segmentation and filtering, with good results in a short time. In this paper, we focus on PET (positron-emission tomography) image denoising with deep convolutional neural networks. Inspired by the idea of encoder-decoder networks, here we propose a simple framework with hierarchical topology of convolu- tional and deconvolutional layers and direct connections between the corresponding hierarchy levels. Our network takes advantage of the a priori structural information included in CT images to better restore details. We show that our method is able to considerably reduce the noise in PET reconstructions. For the sake of completeness, we also analyze the PSNR and SSIM image quality metrics for different experimental settings. LA - English DB - MTMT ER - TY - CONF AU - Csébfalvi, Balázs AU - Rácz, Gergely Ferenc ED - NJSZT, null TI - Optimális BCC rácson mintavételezett 3D jelek folytonos rekonstrukciójának elemzése a frekvenciatartományban T2 - KÉPAF 2017: Képfeldolgozók és Alakfelismerők Társaságának 11. országos konferenciája PB - Neumann János Számítógép-tudományi Társaság C1 - Budapest PY - 2017 PG - 10 UR - https://m2.mtmt.hu/api/publication/3184353 ID - 3184353 LA - English DB - MTMT ER - TY - CHAP AU - Rácz, Gergely Ferenc AU - Csébfalvi, Balázs ED - Kiss, Bálint ED - Szirmay-Kalos, László TI - Designing Interpolation Filters for Isotropic Tomographic Reconstruction T2 - Proceedings of the Workshop on the Advances of Information Technology: WAIT 2017 PB - BME Irányítástechnika és Informatika Tanszék CY - Budapest SN - 9789633132425 PY - 2017 SP - 74 EP - 81 PG - 8 UR - https://m2.mtmt.hu/api/publication/3184326 ID - 3184326 LA - English DB - MTMT ER - TY - JOUR AU - Csébfalvi, Balázs AU - Rácz, Gergely Ferenc TI - Retailoring Box Splines to Lattices for Highly Isotropic Volume Representations JF - COMPUTER GRAPHICS FORUM J2 - COMPUT GRAPH FORUM VL - 35 PY - 2016 IS - 3 SP - 411 EP - 420 PG - 10 SN - 0167-7055 DO - 10.1111/cgf.12917 UR - https://m2.mtmt.hu/api/publication/3120829 ID - 3120829 AB - 3D box splines are defined by convolving a 1D box function with itself along different directions. In volume visualization, box splines are mainly used as reconstruction kernels that are easy to adapt to various sampling lattices, such as the Cartesian Cubic (CC), Body-Centered Cubic (BCC), and Face-Centered Cubic (FCC) lattices. The usual way of tailoring a box spline to a specific lattice is to span the box spline by exactly those principal directions that span the lattice itself. However, in this case, the preferred directions of the box spline and the lattice are the same, amplifying the anisotropic effects of each other. This leads to an anisotropic volume representation with strongly preferred directions. Therefore, in this paper, we retailor box splines to lattices such that the sets of vectors that span the box spline and the lattice are disjoint sets. As the preferred directions of the box spline and the lattice compensate each other, a more isotropic volume representation can be achieved. We demonstrate this by comparing different combinations of box splines and lattices concerning their anisotropic behavior in tomographic reconstruction and volume visualization. LA - English DB - MTMT ER -