Papers by Kamel Belloulata
Bi-ESRGAN: A New Approach of Document Image Super-Resolution Based on Dual Deep Transfer Learning
Communications in Computer and Information Science
Compression Fractale basée-région des séquences vidéo segmentées
ABSTRACT - Nous proposons une nouvelle méthode de compression basée région, des séquences vidéo, ... more ABSTRACT - Nous proposons une nouvelle méthode de compression basée région, des séquences vidéo, basée sur le codage fractal. Chaque objet de la séquence est individuellement codé par un système de fonctions itératifs locales (SFIL) approprié. L'algorithme utilise une prédiction circulaire entre les quatre (4) premières trames de la séquence vidéo (Circular Prédiction Mapping -CPM-) et des transformations affines non contractantes (Non Contractive Interframe Mapping -NCIM-) pour les trames suivantes. Le CPM et NCIM combinent le codage fractal avec l'algorithme d'estimation et de compensation de mouvement (ME/MC), en exploitant la corrélation temporelle entre les trames adjacentes. Pour introduire la fonctionnalité basée - région, chaque région de la séquence doit être codée indépendamment en accord avec une carte de segmentation. Ceci permet une transmission/stockage indépendant et, par conséquent, un décodage individuel des régions. Les blocs (source/cibles) peuvent donc contenir plus qu'un segment. Dans ce scénario, chaque segment dans un bloc cible peut être modélisé par un segment d'un bloc source du même objet. Notre codeur fractal basé - région des images fixes pourra générer des résultats plus intéressants dans le contexte vidéo. Cependant, cet algorithme sera utilisé pour coder des objets définis par une carte de segmentation, exactement comme dans MPEG-4. Il a été prouvé par simulation sur une station de travail que l'algorithme proposé produit des images de qualité visuels intéressantes pour la vidéo conférence à un débit très faible, inférieure à 64 kbps.
Architecture d’authentification pour assurer la mobilité et la qualité de service dans les réseaux sans fil
Reconstruction d'images par la méthode de l'échantillonnage compressif: Application à la fusion d'image satellitaire et à l'IRM
Compression d'images par fractale : études sur la mesure et le domaine de recherche de l'autosimilarité (spatial ou transforme) et sur l'accélération de la génération du modèle fractal
L'objectif de ce travail est le developpement d'une methodologie optimisee de compression... more L'objectif de ce travail est le developpement d'une methodologie optimisee de compression des images fondee sur la theorie des fractales. La compression par fractale est basee sur la detection, la meure et le codage de l'auto-similarite presente dans les scenes naturelles se manifestant en termes de ressemblance entre des structures locales a differentes echelles, orientations et dispositions spatiales. Apres un partitionnement adapte de l'image, chaque partie elementaire est mise en correspondance avec une autre partie d'echelle differente recherchee dans toute l'image. Chaque correspondance est modelisee par une transformation affine locale et l'union de ces transformations appelee Local Iterated Funcion Systems (LIFS) forme le code fractal. En ce qui concerne le schema de base, des ameliorations sont apportee s a differents niveaux. L'influence du choix de la norme utilisee dans la mesure de l'auto-similarite est etudiee et une nouvelle methode de codage base sur la norme L est proposee et comparee a celle basee sur a norme Lα. Afin de reduire la complexite du codage par fractale des images pleine resolution, une classification parametrique des blocs source et destination est faite en utilisant la variance ou la dimension fractales des blocs cible. En ce qui concerne les schemas hybrides (fractale/TCD et fractale / sous bandes), une synthese complete de l'etat de l'art est effectuee. Un e methode originale de codage par fractale intra-sou s bandes adaptatif est proposee, methode qui permet de profiter des avantages des deux theories (ondelettes et fractale) pour augmenter les performances de l'algorithme de codage en termes de minimisation de la distorsion et du temps de calcul. En ce qui concerne l'acceleration du codage par fractale, un nouvel algorithme de compression hybride fractale/QV/sous-bandes avec une approche non-iterative de classification des blocs source est developpe et compare aux algorithmes d'acceleration existants en termes de complexite d'algorithme et temps de codage.
Region-based fractal video coding
Proceedings of SPIE, Oct 10, 2008
ABSTRACT A novel video sequence compression scheme is proposed in order to realize the efficient ... more ABSTRACT A novel video sequence compression scheme is proposed in order to realize the efficient and economical transmission of video sequence, and also the region-based functionality of MPEG-4. The CPM and NCIM fractal coding scheme is applied on each region independently by a prior image segmentation map (alpha plane) which is exactly the same as defined in MPEG-4. The first n frames of video sequence are encoded as a "set" using the Circular Prediction Mapping (CPM) and encode the remaining frames using the Non Contractive Interframe Mapping (NCIM). The CPM and NCIM accomplish the motion estimation and compensation, which can exploit the high temporal correlations between the adjacent frames of video sequence. The experimental results with the monocular video sequences provide promising performances at low bit rate coding, such as the application in video conference. We believe the proposed fractal video codec will be a powerful and efficient technique for the region-based video sequence coding.
A Hybrid Video Coding Scheme Joint Fractal and Object-Based Function
ABSTRACT Based on the classical fractal video coding scheme, a novel hybrid video coding scheme j... more ABSTRACT Based on the classical fractal video coding scheme, a novel hybrid video coding scheme joint fractal and object-based function is proposed. A new technique of object-based fractal video coding in which each object can be encoded and decoded independently results in increasing compression ratio and speed and reducing bit rate and bandwidth is used. The proposed techniques include: using more effective macroblock partition scheme instead of classical quadtree partition scheme; reducing the block searching strategy and range, thus increasing the calculation speed greatly; using homo-I-frame like in H.264; reducing repeating calculation, etc. Experimental results indicate that the proposed method can raise compression ratio 4 times, speed up compression time 10 times, and improve the image quality 3 to 5 dB, the proposed object-based scheme improves the performances of fractal video coding algorithm greatly, and the hybrid object-based fractal video coding scheme is simple and effective, it adds more flexibility and practicability of fractal video coding applications.
A novel fractal video coding algorithm using fast block-matching motion estimation technology
ABSTRACT We proposed a novel fractal video coding method using fast block-matching motion estimat... more ABSTRACT We proposed a novel fractal video coding method using fast block-matching motion estimation technology to overcome the drawback of the time-consuming character in the fractal coding. As fractal encoding essentially spends most time on the search for the best-matching block in a large domain pool, this paper firstly ameliorates the traditional CPM/NCIM method and applies a new hexagon block-matching motion estimation technology into the fractal video coding. Experimental results indicate that the proposed algorithm spends less encoding time and achieves higher compression ratio and the compression quality compared with the traditional CPM/NCIM method.
An improved fractal image coding algorithm based on adaptive threshold for quadtree partition
Proceedings of SPIE, Oct 10, 2008
ABSTRACT The image coding algorithm based on quadtree partitioning, which is presented by Yuval F... more ABSTRACT The image coding algorithm based on quadtree partitioning, which is presented by Yuval Fisher, has become the mainstream of fractal image compression. However, in the conventional Yuval Fisher's theory, the threshold of distance between range blocks and domain blocks is pre-fixed mainly by experience, which does not take the practical statistical characteristics of range blocks into account for automatic valuation. Thus the coding algorithm could not adapt to various input images well, and sometimes the performances on certain images are not further optimized. To overcome this shortcoming, an improved adaptive threshold (AT) for quadtree partition is presented in this paper. Compared with the fixed threshold (FT) partition, this improved method could shorten coding time and obtain higher PSNR that have been demonstrated by the experimental results. Furthermore, the decoded images get more subjective acceptances in accordance with the standard of Human Visual System (HVS).
Applied Mathematical Modelling, Nov 1, 2012
In this paper, we explore the fractal video sequences coding in the context of region-based funct... more In this paper, we explore the fractal video sequences coding in the context of region-based functionality. Since the main drawback of fractal coding is the high computational complexity, some schemes are proposed to speed up the encoding process. As fractal encoding essentially spends most time on the search for the best-matching block in a large domain pool, this paper firstly ameliorates the conventional CPM/NCIM method and then applies a new hexagon block-matching motion estimation technology into the fractal video coding. The images in the video sequences are encoded region by region according to a previouslycomputed segmentation map. Experimental results indicate that the proposed algorithm spends less encoding time and achieves higher compression ratio and compression quality compared with the conventional CPM/NCIM method.
An Image Segmentation Algorithm in Image Processing Based on Threshold Segmentation
Image segmentation is a key technology in image processing, and threshold segmentation is one of ... more Image segmentation is a key technology in image processing, and threshold segmentation is one of the methods used frequently. Aimed at that only one threshold or several thresholds are set in traditional threshold-based segmentation algorithm, it is difficult to extract the complex information in an image, a new segmentation algorithm that each pixel in the image has its own threshold
A novel spatio-temporal video object segmentation algorithm
ABSTRACT Video object segmentation is a key technology in video processing, which is the foundati... more ABSTRACT Video object segmentation is a key technology in video processing, which is the foundation of the newly object-based video compression standard MPEG-4. A novel spatio-temporal video object segmentation algorithm is proposed in this paper, which is very efficient to acquire the video object moving in a static background. Firstly, video sequence is segmented with change detection algorithm to get segmentation result in temporal domain. Secondly, the spatial segmentation result is obtained by a segmentation algorithm based on local thresholds, which means each pixel in the image has its own threshold. In this algorithm, the threshold of a pixel in an image is estimated by calculating the mean of the grayscale values of its neighbor pixels, and the square variance of the grayscale values of the neighbor pixels is also calculated as an additional judge condition. Thirdly, the video object is extracted by spatio-temporal integration video segmentation algorithm. The experimental results show that the proposed spatio-temporal integration video object segmentation algorithm can extract the moving object in a video sequence with a static background accurately and efficiently.
An automatic region-based video sequence codec based on fractal compression
Aeu-international Journal of Electronics and Communications, Aug 1, 2014
ABSTRACT In order to improve the performance of fractal video coding, we explore a novel fractal ... more ABSTRACT In order to improve the performance of fractal video coding, we explore a novel fractal video sequences codec with automatic region-based functionality. To increase the quality of decoding image, intra frame coding, deblocking loop filter and sub-pixel block matching are applied to the codec. An efficient searching algorithm is used to increase the compression ratio and encoding speed. Automatic region-based fractal video sequences coding reduces coding stream greatly. Experimental results indicate that the proposed algorithm is more robust, and provides much less encoding time and bitrate while maintaining the quality of decompression image than the conventional CPM/NCIM method and other related references. We compare the proposed algorithm with three algorithms in [24–26], and the results of all these four algorithms are compared with H.264. The bitrate of the proposed algorithm is decreased by 0.11% and the other algorithms are increased by 4.29%, 6.85% and 11.62%, respectively. The average PSNR degradations of the four algorithms are 0.71 dB, 0.48 dB, 0.48 dB and 0.75 dB. So the bitrate of the proposed algorithm is decreased and the other algorithms are increased. At the meantime the compression time is reduced greatly, about 79.19% on average. The results indicate that, on average, the proposed automatic region-based fractal video sequences coding system can save compression time 48.97% and bitrate 52.02% with some image quality degradation in comparison with H.264, since they are all above 32 dB and the human eyes are insensitive to the differences.
ISRN Signal Processing (Print), Jul 12, 2011
We propose a novel fractal video coding method using fast block-matching motion estimation to ove... more We propose a novel fractal video coding method using fast block-matching motion estimation to overcome the drawback of the time-consuming character in the fractal coding. As fractal encoding essentially spends most time on the search for the bestmatching block in a large domain pool, search patterns and the center-biased characteristics of motion vector distribution have large impact on both search speed and quality of block motion estimation. In this paper, firstly, we propose a new hexagon search algorithm (NHEXS), and, secondly, we ameliorate, by using this NHEXS, the traditional CPM/NCIM, which is based on Fisher's quadtree partition. This NHEXS uses two cross-shaped search patterns as the first two initial steps and large/small hexagon-shaped patterns as the subsequent steps for fast block motion estimation (BME). NHEXS employs halfway stop technique to achieve significant speedup on sequences with stationary and quasistationary blocks. To further reduce the computational complexity, NHEXS employs modified partial distortion criterion (MPDC). Experimental results indicate that the proposed algorithm spends less encoding time and achieves higher compression ratio and compression quality compared with the traditional CPM/NCIM method.
Indexation Basée-Région d’Image Utilisant Shape-Adaptive DCT (SA-DCT)
HAL (Le Centre pour la Communication Scientifique Directe), Dec 2, 2015
International audienceCe papier propose une nouvelle approche dans le domaine de la recherche d’i... more International audienceCe papier propose une nouvelle approche dans le domaine de la recherche d’images par région (RBIR), utilisant la transformation cosinus discrète adaptée aux objets (SA-DCT). Dans ce système de recherche, une image a une segmentation a priori (alpha-plane), tel qu’il a était définit dans MPEG-4. Autrement dit, une image est représentée par des régions segmentées, et chaque région est associée à un vecteur de caractéristiques dérivé à partir de la DCT et la SA-DCT. Une fois les caractéristiques sont extraites, on désigne deux systèmes différents pour tester l’efficacité de la recherche. Ces deux systèmes diffèrent dans la formulation de la requête et le calcul de la mesure de similarité entre les images. Le premier système utilise la région d’intérêt comme requête, on le nomme: Technique basée sur une seule région (région par région) . Dans ce système, la distance entre la requête et les images de la base de données est le minimum de distances entre la région requête et toutes les régions des images de la base de données. Le deuxième système utilise une requête globale dans lequel une image requête entière est comparée avec les images entières de la base de données. On nomme ce système: Technique basée sur l’image globale (image par image). Les distances individuelles entre les régions des deux images sont combinées pour avoir une seule mesure de distance.Les résultats obtenus montrent que l’approche proposée est capable d’identifier l’objet principal et de réduire l’influence de l’arrière-plan, ceci améliore la performance de la recherche d’images en comparaison avec d’autres systèmes conventionnels utilisant la DCT dans le domaine CBIR et même avec des systèmes dans le domaine RBI
Lecture Notes in Computer Science, 2009
A novel object-based fractal monocular and stereo video compression scheme with quadtree-based mo... more A novel object-based fractal monocular and stereo video compression scheme with quadtree-based motion and disparity compensation is proposed in this paper. Fractal coding is adopted and each object is encoded independently by a prior image segmentation alpha plane, which is defined exactly as in MPEG-4. The first n frames of right video sequence are encoded by using the Circular Prediction Mapping (CPM) and the remaining frames are encoded by using the Non Contractive Interframe Mapping (NCIM). The CPM and NCIM methods accomplish the motion estimation/compensation of right video sequence. According to the different coding or user requirements, the spatial correlations between the left and right frames can be explored by partial or full affine transformation quadtree-based disparity estimation/compensation, or simply by applying CPM/NCIM on left video sequence. The testing results with monocular and stereo video sequences provide promising performances at low bit rate coding. We believe it will be a powerful and efficient technique for the object-based monocular and stereo video sequences coding.
A Fractal Image Coding Scheme Based on Norm
Turkish Journal of Electrical Engineering and Computer Sciences, 1997
Information, Dec 22, 2022
This article is an open access article distributed under the terms and conditions of the Creative... more This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY
A New Object-Based Fractal Stereo Codec with Quadtree-Based Disparity or Motion Compensation
ABSTRACT A new object-based fractal stereo video coding scheme with quadtree-based disparity or m... more ABSTRACT A new object-based fractal stereo video coding scheme with quadtree-based disparity or motion compensation is proposed in this paper. Monocular or stereo video are compressed by using fractal coding and each object is encoded independently by a prior image segmentation alpha plane which is exactly the same as in MPEG-4. The first 2 frames of right video are encoded by using the circular prediction mapping (CPM) and the remaining frames are encoded by using the non contractive interframe mapping (NCIM). The CPM and NCIM methods accomplish the motion estimation/compensation of right video (reference), and according to different coding or user requirements, the spatial correlations between left and right frames can be explored by partial or full affine transformation quadtree-based disparity estimation/compensation, or simply applying CPM/NCIM on left video. The experimental results indicate that we get high compression ratio and also good PSNR
Journal of Multimedia, Jun 1, 2007
A novel object-based fractal monocular and stereo video compression scheme with quadtreebased mot... more A novel object-based fractal monocular and stereo video compression scheme with quadtreebased motion and disparity compensation is proposed in this paper. Fractal coding is adopted and each object is encoded independently by a prior image segmentation alpha plane, which is defined exactly as in MPEG-4. The first n frames of right video sequence are encoded by using the Circular Prediction Mapping (CPM) and the remaining frames are encoded by using the Non Contractive Interframe Mapping (NCIM). The CPM and NCIM methods accomplish the motion estimation/compensation of right video sequence. According to the different coding or user requirements, the spatial correlations between the left and right frames can be explored by partial or full affine transformation quadtree-based disparity estimation/ compensation, or simply by applying CPM/NCIM on left video sequence.
Uploads
Papers by Kamel Belloulata