Mostrar el registro sencillo del documento

dc.rights.licenseReconocimiento 4.0 Internacional
dc.contributor.advisorBranch Bedoya, John Willian (Thesis advisor)
dc.contributor.advisorTrujillo Uribe, Maria Patricia
dc.contributor.authorSalazar Herrera, Carlos Alberto
dc.date.accessioned2024-07-10T13:57:27Z
dc.date.available2024-07-10T13:57:27Z
dc.date.issued2023-07-09
dc.identifier.urihttps://repositorio.unal.edu.co/handle/unal/86427
dc.descriptionIlustraciones
dc.description.abstractVideo in-loop restoration methods have traction much attention across the standardization groups for future video codecs AV2 1 and VVC 2. Primarily, because of potential benefits to compensate the artifacts generated during super-resolution scenarios and the effect of quan- tization process. Thus, new sophisticated learned-based algorithms have been proposed, in recent years, surpassing the classical switchable filter implementation in objective quality rate. However, CNN-based approaches requirements on computational cost and decoder complexity are still challenging. Therefore, we propose a low-complex learned-based method that leverages the solid and consistent sparse representation theory to exploit the spatial redundancy of frames. Our approach models the decoding residual, the distance between each reference and the respec- tive decoded frame. Furthermore, the proposed methods shrink the support of the sparse vector to two in order to control the restoration signal information. In addition, our method uses the Discrete Cosine Transform (DCT) orthogonal basis as a dictionary to exploit the statistical correlation between nonzero coefficients and the quantization level. Finally, we leverage the official and public available AV2 raw video dataset to compare our performance against the anchor AV2 codec through three objective visual quality metrics. The validation protocol includes benchmark data sets for the anchor and the restoration-enabled configurations. Our experimental results show a consistent restoration using sparse representation as well as an effective mechanism for sharing nonzero coefficients leveraging a Gaussian correlation. The experimental evaluation showed that our method has a 1%-2% gain regarding AV2, using SSIM and VMAF under similar bitrate conditions.
dc.description.abstractLos métodos de restauración de vídeo en bucle han venido incrementando el intereste por parte los grupos de estandarización para los futuros códecs de vídeo AV2 y VVC. Esto principalmente debido a a los beneficios potenciales para compensar efectos no deseados en el video producidos durante los procesos de super-resolución y cuantización. Así, en los últimos años se han propuesto nuevos y sofisticados algoritmos basados en aprendizaje, que superan a la clásica implementación de filtros conmutables en cuanto a tasa de calidad objetiva. Sin embargo, los requisitos de los enfoques basados en CNN en cuanto a coste computacional y complejidad del descodificador siguen siendo un desafio. Por ello, proponemos un método de baja complejidad basado en aprendizaje, que aprovecha la sólida y consistente teoría de la representación dispersa para explotar la redundancia espacial de los fotogramas que componen un video. Nuestro enfoque modela el residuo de descodificación, la distancia entre cada referencia y el respectivo fotograma descodificado. Además, el método propuesto reduce el soporte del vector disperso a dos para controlar la información de la señal de restauración. Por otra parte, nuestro método utiliza la base ortogonal de la transformada discreta de coseno (DCT) como diccionario para explotar la correlación estadística entre los coeficientes distintos de cero y el nivel de cuantificación. Por último, aprovechamos el conjunto de datos de vídeo de AV2, oficial y público, para comparar nuestro rendimiento con el códec AV2 de referencia, mediante tres métricas objetivas de calidad visual.El protocolo de validación incluye conjuntos de datos de referencia para las funciones de anclaje y restauración. Nuestros resultados experimentales muestran una restauración coherente utilizando una representación dispersa así como un mecanismo eficaz para compartir coeficientes distintos de cero aprovechando una correlación gaussiana. La evaluación experimental mostró que nuestro método tiene una ganancia del 1%-2% con respecto a AV2, utilizando SSIM y VMAF en condiciones de bitrate similares.
dc.format.extent74 páginas
dc.format.mimetypeapplication/pdf
dc.language.isoeng
dc.publisherUniversidad Nacional de Colombia
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.subject.ddc000 - Ciencias de la computación, información y obras generales::003 - Sistemas
dc.subject.ddc000 - Ciencias de la computación, información y obras generales::004 - Procesamiento de datos Ciencia de los computadores
dc.titleSparse In-loop Video Coding Restoration Method
dc.typeTrabajo de grado - Doctorado
dc.type.driverinfo:eu-repo/semantics/doctoralThesis
dc.type.versioninfo:eu-repo/semantics/acceptedVersion
dc.publisher.programMedellín - Minas - Doctorado en Ingeniería - Sistemas
dc.contributor.researchgroupGidia: Grupo de Investigación YyDesarrollo en Inteligencia Artificial
dc.contributor.subjectmatterexpertTrujillo Uribe, Maria Patricia
dc.description.degreelevelDoctorado
dc.description.degreenameDoctor en Ingeniería
dc.description.researchareaAnálisis de video
dc.identifier.instnameUniversidad Nacional de Colombia
dc.identifier.reponameRepositorio Institucional Universidad Nacional de Colombia
dc.identifier.repourlhttps://repositorio.unal.edu.co/
dc.publisher.facultyFacultad de Minas
dc.publisher.placeMedellín, Colombia
dc.publisher.branchUniversidad Nacional de Colombia - Sede Medellín
dc.relation.referencesSparse Representations. 2009
dc.relation.referencesAhmed, N. ; Natarajan, T. ; Rao, K.R.: Discrete Cosine Transform. En: IEEE Transactions on Computers C-23 (1974), Nr. 1, p. 90–93
dc.relation.referencesAntsiferova, Anastasia ; Lavrushkin, Sergey ; Smirnov, Maksim ; Gushchin, Aleksandr ; Vatolin, Dmitriy S. ; Kulikov, Dmitriy. Video compression dataset and benchmark of learning-based video-quality metrics. 2022
dc.relation.referencesBarman, Nabajeet ; Martini, Maria G. ; Reznik, Yuriy: Revisiting Bjontegaard Delta Bitrate (BD-BR) Computation for Codec Compression Efficiency Comparison. En: Proceedings of the 1st Mile-High Video Conference. New York, NY, USA : Association for Computing Machinery, 2022 (MHV ’22). – ISBN 9781450392228, p. 113–114
dc.relation.referencesCai, T. T. ; Wang, Lie: Orthogonal matching pursuit for sparse signal recovery with noise. En: IEEE Transactions on Information Theory 57 (2011), 7, p. 4680–4688. –ISSN 00189448
dc.relation.referencesChen, Ching-Yeh ; Tsai, Chia-Yang ; Huang, Yu-Wen ; Yamakage, Tomoo ; Chong, In S. ; Fu, Chih-Ming ; Itoh, Takayuki ; Watanabe, Takashi ; Chujoh, Takeshi ; Karczewicz, Marta ; Lei, Shaw-Min: The adaptive loop filtering techniques in the HEVC standard. En: Tescher, Andrew G. (Ed.): Applications of Digital Image Processing XXXV Vol. 8499, 2012, p. 849913
dc.relation.referencesDai, Yuanying ; Liu, Dong ; Wu, Feng: A Convolutional Neural Network Approach for Post-Processing in HEVC Intra Coding
dc.relation.referencesDing, Dandan ; Chen, Guangyao ; Mukherjee, Debargha ; Joshi, Urvang ; Chen, Yue: A CNN-based In-loop Filtering Approach for AV1 Video Codec. En: 2019 Picture Coding Symposium (PCS), 2019, p. 1–5
dc.relation.referencesDIng, Dandan ; Chen, Guangyao ; Mukherjee, Debargha ; Joshi, Urvang ; Chen, Yue: A progressive CNN in-loop filtering approach for inter frame coding. En: 2019 Picture Coding Symposium, PCS 2019 (2019). ISBN 9781728147048
dc.relation.referencesDong, Junshuo ; Wu, Lingda: Comparison and Simulation Study of the Sparse Representation Matching Pursuit Algorithm and the Orthogonal Matching Pursuit Algorithm. (2021), p. 317–320
dc.relation.referencesDong, Weisheng ; Shi, Guangming ; Li, Xin: Image deblurring with low-rank approximation structured sparse representation. En: 2012 Conference Handbook - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, AP-SIPA ASC 2012 (2012), p. 14–18. ISBN 9780615700502
dc.relation.referencesDong, Weisheng ; Zhang, Lei ; Shi, Guangming ; Li, Xin: Nonlocally centralized sparse representation for image restoration. En: IEEE Transactions on Image Processing 22 (2013), Nr. 4, p. 1620–1630. – ISSN 10577149
dc.relation.referencesDong, Weisheng ; Zhang, Lei ; Shi, Guangming ; Li, Xin: Nonlocally centralized sparse representation for image restoration. En: IEEE Transactions on Image Processing 22 (2013), Nr. 4, p. 1620–1630. – ISSN 10577149
dc.relation.referencesDong, Weisheng ; Zhang, Lei ; Shi, Guangming ; Wu, Xiaolin: Image deblurring and super-resolution by adaptive sparse domain selection and adaptive regularization. En: IEEE Transactions on Image Processing 20 (2011), Nr. 7, p. 1838–1857. – ISSN 10577149
dc.relation.referencesDonoho, David L.: Compressed sensing. En: IEEE Transactions on Information Theory 52 (2006), Nr. 4, p. 1289–1306. – ISSN 00189448
dc.relation.referencesDonoho, David L. ; Elad, Michael: Optimally sparse representation in general (nonorthogonal) dictionaries via 1 minimization. En: PNAS March 4 (2003), p. 2197–2202
dc.relation.referencesEbadi, Salehe E. ; Ones, Valia G. ; Izquierdo, Ebroul: UHD Video Super-Resolution Using Low-Rank and Sparse Decomposition. En: Proceedings - 2017 IEEE International Conference on Computer Vision Workshops, ICCVW 2017 2018-Janua (2017), p. 1889–1897. ISBN 9781538610343
dc.relation.referencesEkstrom, Michael P.: Realizable Wiener Filtering in Two Dimensions. En: IEEE Transactions on Acoustics, Speech, and Signal Processing 30 (1982), p. 31–40. – ISSN 00963518
dc.relation.referencesElad, Michael ; Aharon, Michal: Image denoising via learned dictionaries and sparse representation. En: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition 1 (2006), p. 895–900. – ISBN 0769525970
dc.relation.referencesElad, Michael ; Bruckstein, Alfred M.: A generalized uncertainty principle and sparse representation in pairs of bases. En: IEEE Transactions on Information Theory 48 (2002), 9, p. 2558–2567. – ISSN 00189448
dc.relation.referencesHan, Jingning ; Li, Bohan ; Mukherjee, Debargha ; Chiang, Ching-Han ; Chen, Cheng ; Su, Hui ; Parker, Sarah ; Joshi, Urvang ; Chen, Yue ; Wang, Yunqing ; Wilkins, Paul ; Xu, Yaowu ; Bankoski, James: A Technical Overview of AV1. (2020), 8
dc.relation.referencesHan, Jingning ; Xu, Yaowu ; Mukherjee, Debargha: A butterfly structured design of the hybrid transform coding scheme. (2013), p. 17–20
dc.relation.referencesHastie, Trevor ; Martin, Robert T. ; Hastie, Wainwright ; Tibshirani, * ; Wainwright, *. Statistical Learning with Sparsity The Lasso and Generalizations Statistical Learning with Sparsity
dc.relation.referencesJi, Hui ; Huang, Sibin ; Shen, Zuowei ; Xu, Yuhong: Robust Video Restoration by Joint Sparse and Low Rank Matrix Approximation. En: SIAM Journal on Imaging Sciences 4 (2011), Nr. 4, p. 1122–1142
dc.relation.referencesJia, Chuanmin ; Wang, Shiqi ; Zhang, Xinfeng ; Wang, Shanshe ; Liu, Jiaying ; Pu, Shiliang ; Ma, Siwei: Content-Aware Convolutional Neural Network for In-Loop Filtering in High Efficiency Video Coding. En: IEEE Transactions on Image Processing 28 (2019), p. 3343–3356. – ISSN 19410042
dc.relation.referencesKato, Toshiyuki ; Hino, Hideitsu ; Murata, Noboru: Sparse Coding Approach for Multi-Frame Image Super Resolution. (2014), p. 1–20
dc.relation.referencesKim, Jiwon ; Lee, Jung K. ; Lee, Kyoung M.: Accurate Image Super-Resolution Using Very Deep Convolutional Networks. En: CoRR abs/1511.04587 (2015)
dc.relation.referencesKong, Lingyi ; Ding, Dandan ; Liu, Fuchang ; Mukherjee, Debargha ; Joshi, Urvang ; Chen, Yue: Guided CNN Restoration with Explicitly Signaled Linear Combination. En: Proceedings - International Conference on Image Processing, ICIP 2020- Octob (2020), p. 3379–3383. – ISBN 9781728163956
dc.relation.referencesLin, Liqun ; Yu, Shiqi ; Zhao, Tiesong ; Wang, Zhou: PEA265: Perceptual Assessment of Video Compression Artifacts. (2019)
dc.relation.referencesMackiewicz, Andrzej ; Ratajczak, Waldemar: Principal Components Analysis (PCA). En: Computers & Geosciences 19 (1993), p. 303–342
dc.relation.referencesMairal, Julien ; Elad, Michael ; Sapiro, Guillermo: Sparse representation for color image restoration. En: IEEE Transactions on Image Processing 17 (2008), p. 53–69. ISSN 10577149
dc.relation.referencesMairal, Julien ; Sapiro, Guillermo ; Elad, Michael: Learning multiscale sparse representations for image and video restoration. En: Multiscale Modeling and Simulation (2008), Nr. 1, p. 214–241. – ISSN 15403467
dc.relation.referencesMoorthy, Anush K. ; Bovik, Alan C. STATISTICS OF NATURAL IMAGE DISTORTIONS
dc.relation.referencesMukherjee, Debargha ; Han, Jingning ; Bankoski, Jim ; Bultje, Ronald ; Grange, Adrian ; Koleszar, John ; Wilkins, Paul ; Xu, Yaowu: A Technical Overview of VP9 – The Latest Open-Source Video Codec. (2013), p. 1–17
dc.relation.referencesO’Shea, Keiron ; Nash, Ryan: An Introduction to Convolutional Neural Networks. En: CoRR abs/1511.08458 (2015)
dc.relation.referencesOxford: A Dictionary of Statistics. Oxford University Press, 2014. – ISBN 9780191758317
dc.relation.referencesReininger, Randall C. ; Gibson, Jerry D.: Distributions of the Two-Dimensional DCT Coefficients for Images. En: IEEE Transactions on Communications 31 (1983), p. 835–839. – ISSN 00906778
dc.relation.referencesSaad, Michele A. ; Bovik, Alan C. ; Charrier, Christophe: DCT statistics model- based blind image quality assessment, 2011. – ISBN 9781457713033, p. 3093–3096
dc.relation.referencesSankaraiah, Yediga R. ; Varadarajan, Sourirajan: An effective image deblurring scheme using cluster based sparse representation. En: ASEAN Engineering Journal 11 (2021), Nr. 4, p. 16–28. – ISSN 25869159
dc.relation.referencesScetbon, Meyer ; Elad, Michael ; Milanfar, Peyman: Deep K-SVD denoising. En: IEEE Transactions on Image Processing 30 (2021), Nr. 8, p. 5944–5955. – ISSN 19410042
dc.relation.referencesSchneider, Jens ; Sauer, Johannes ; Wien, Mathias: RDPlot – An Evaluation Tool for Video Coding Simulations. En: 2021 International Conference on Visual Communications and Image Processing (VCIP), 2021, p. 1–1
dc.relation.referencesSegall, C A. ; Katsaggelos, Aggelos K. ; Molina, Rafael: Chapter 11 Super- resolution from compressed video. En: Book (2001), p. 1–32
dc.relation.referencesSiekmann, Mischa ; Bosse, Sebastian ; Schwarz, Heiko ; Wiegand, Thomas: SEPARABLE WIENER FILTER BASED ADAPTIVE IN-LOOP FILTER FOR VIDEO CODING Image Processing Department Fraunhofer Institute for Telecommunications
dc.relation.referencesValin, Jean-Marc: The Daala Directional Deringing Filter. En: CoRR abs/1602.05975 (2016)
dc.relation.referencesWang, Z. ; Simoncelli, E.P. ; Bovik, A.C.: Multiscale structural similarity for image quality assessment. En: The Thrity-Seventh Asilomar Conference on Signals, Systems and Computers, 2003 Vol. 2, 2003, p. 1398–1402 Vol.2
dc.relation.referencesWang, Zhou ; Bovik, A.C. ; Sheikh, H.R. ; Simoncelli, E.P.: Image quality assessment: from error visibility to structural similarity. En: IEEE Transactions on Image Processing 13 (2004), Nr. 4, p. 600–612
dc.relation.referencesWiener, Norbert: Extrapolation, interpolation, and smoothing of stationary time series with engineering applications. 1964. – ISBN 9780262730051
dc.relation.referencesY., Dodge: Gamma Distribution. New York, NY : Springer New York, 2008. – 215–216 p.. – ISBN 978–0–387–32833–1
dc.relation.referencesYang, Jianchao ; Wright, John ; Huang, Thomas S. ; Ma, Yi: Image super- resolution via sparse representation. En: IEEE Transactions on Image Processing 19 (2010), Nr. 11, p. 2861–2873. – ISSN 10577149
dc.relation.referencesZhu, Shujin ; Yu, Zekuan: Self-guided filter for image denoising. En: IET Image Processing 14 (2020), Nr. 11, p. 2561–2566
dc.rights.accessrightsinfo:eu-repo/semantics/openAccess
dc.subject.lembVideos - Conservación y restauración
dc.subject.lembProcesamiento digital de imágenes
dc.subject.proposalVideo compression
dc.subject.proposalrestoration
dc.subject.proposalsparse
dc.subject.proposalAV2
dc.subject.proposalHEVC
dc.subject.proposalVVC
dc.subject.proposalQP
dc.title.translatedMétodo para la restauración de video en el bucle del proceso de compresión
dc.type.coarhttp://purl.org/coar/resource_type/c_db06
dc.type.coarversionhttp://purl.org/coar/version/c_ab4af688f83e57aa
dc.type.contentText
dc.type.redcolhttp://purl.org/redcol/resource_type/TD
oaire.accessrightshttp://purl.org/coar/access_right/c_abf2
dcterms.audience.professionaldevelopmentInvestigadores
dc.description.curricularareaÁrea Curricular de Ingeniería de Sistemas e Informática
dc.contributor.orcidSalazar Herrera, arlos Alberto [0000000194098229]
dc.contributor.cvlacSALAZAR HERRERA, CARLOS ALBERTO


Archivos en el documento

Thumbnail

Este documento aparece en la(s) siguiente(s) colección(ones)

Mostrar el registro sencillo del documento

Reconocimiento 4.0 InternacionalEsta obra está bajo licencia internacional Creative Commons Reconocimiento-NoComercial 4.0.Este documento ha sido depositado por parte de el(los) autor(es) bajo la siguiente constancia de depósito