Cette thèse propose d'explorer des méthodes de représentations parcimonieuses et d'apprentissage de dictionnaires pour compresser et classifier des images satellites. Les représentations parcimonieuses consistent à approximer un signal par une combinaison linéaire de quelques colonnes, dites atomes, d'un dictionnaire, et ainsi à le représenter par seulement quelques coefficients non nuls contenus dans un vecteur parcimonieux. Afin d'améliorer la qualité des représentations et d'en augmenter la parcimonie, il est intéressant d'apprendre le dictionnaire. La première partie de la thèse présente un état de l'art consacré aux représentations parcimonieuses et aux méthodes d'apprentissage de dictionnaires. Diverses applications de ces méthodes y sont détaillées. Des standards de compression d'images sont également présentés. La deuxième partie traite de l'apprentissage de dictionnaires structurés sur plusieurs niveaux, d'une structure en arbre à une structure adaptative, et de leur application au cas de la compression d'images satellites en les intégrant dans un schéma de codage adapté. Enfin, la troisième partie est consacrée à l'utilisation des dictionnaires structurés appris pour la classification d'images satellites. Une méthode pour estimer la Fonction de Transfert de Modulation (FTM) de l'instrument dont provient une image est étudiée. Puis un algorithme de classification supervisée, utilisant des dictionnaires structurés rendus discriminants entre les classes à l'apprentissage, est présenté dans le cadre de la reconnaissance de scènes au sein d'une image. / This thesis explores sparse representation and dictionary learning methods to compress and classify satellite images. Sparse representations consist in approximating a signal by a linear combination of a few columns, known as atoms, from a dictionary, and thus representing it by only a few non-zero coefficients contained in a sparse vector. In order to improve the quality of the representations and to increase their sparsity, it is interesting to learn the dictionary. The first part of the thesis presents a state of the art about sparse representations and dictionary learning methods. Several applications of these methods are explored. Some image compression standards are also presented. The second part deals with the learning of dictionaries structured in several levels, from a tree structure to an adaptive structure, and their application to the compression of satellite images, by integrating them in an adapted coding scheme. Finally, the third part is about the use of learned structured dictionaries for the classification of satellite images. A method to estimate the Modulation Transfer Function (MTF) of the instrument used to capture an image is studied. A supervised classification algorithm, using structured dictionaries made discriminant between classes during the learning, is then presented in the scope of scene recognition in a picture.

No description available.

No description available.

No description available.

Les technologies numériques et du multimédia ont connu de grandes avancées ces dernières années. La chaîne de transmission des images est constituée de plusieurs traitements divers et variés permettant de transmettre un flux de données toujours plus grand avec toujours plus de services à la clé. Nous citons par exemple, la compression, l'augmentation de contenu, la confidentialité, l'intégrité et l'authenticité des images pendant leur transmission. Dans ce contexte, les approches conjointes ont suscité un intérêt certain de la part de la communauté du traitement d'images car elles permettent d'obtenir des systèmes de faible complexité calculatoire pouvant être utilisés dans des applications nécessitant peu de ressources matérielles. La dissimulation de données ou Data Hiding, est l'art de cacher un message dans un support numérique. L'une des branches les plus importantes du data hiding est le tatouage numérique ou watermarking. La marque doit rester présente dans l'image hôte même si celle-ci subit des modifications appelées attaques. La compression d'images a comme objectif de réduire la taille des images stockées et transmises afin d'augmenter la capacité de stockage et de minimiser le temps de transmission. La compression représente une opération incontournable du stockage ou du transfert d'images. Elle est considérée par le data hiding comme une attaque particulièrement destructrice. La norme JPEG2000 est le dernier standard ISO/ITU-T pour le codage des images fixes. Dans cette thèse, nous étudions de manière conjointe la compression avec perte et le data hiding dans le domaine JPEG2000. L'approche conjointe offre de nombreux avantages dont le plus important est que la compression ne constitue plus une attaque vis-à-vis du data hiding. Les contraintes à respecter sont exprimées en termes de compromis à atteindre: compromis entre la quantité d'information insérée (payload), le taux de compression, la distorsion induite par l'insertion du message et la robustesse de la marque dans le cas du tatouage.Nos travaux de recherche ont conduit à l'élaboration de plusieurs schémas conjoints : un schéma conjoint d'insertion de données cachées et deux schémas conjoints de tatouage dans JPEG2000. Tous ces systèmes conjoints reposent sur des stratégies d'insertion informée basées sur la quantification codée par treillis (TCQ). Les propriétés de codage de canal de la TCQ sont exploitées pour pouvoir à la fois quantifier et insérer un message caché (ou une marque) pendant l'étape de quantification de JPEG2000. / Technological advances in the fields of telecommunications and multimedia during the two last decades, derive to create novel image processing services such as copyright protection, data enrichment and information hiding applications. There is a strong need of low complexity applications to perform seveval image processing services within a single system. In this context, the design of joint systems have attracted researchers during the last past years. Data hiding techniques embed an invisible message within a multimedia content by modifying the media data. This process is done in such a way that the hidden data is not perceptible to an observer. Digital watermarking is one type of data hiding. The watermark should be resistant to a variety of manipulations called attacks. The purpose of image compression is to represent images with less data in order to save storage costs or transmission time. Compression is generally unavoidable for transmission or storage purposes and is considered as one of the most destructive attacks by the data hiding. JPEG2000 is the last ISO/ ITU-T standard for still image compression.In this thesis, joint compression and data hiding is investigated in the JPEG2000 framework. Instead of treating data hiding and compression separately, it is interesting and beneficial to look at the joint design of data hiding and compression system. The joint approach have many advantages. The most important thing is that compression is no longer considered as an attack by data hiding.The main constraints that must be considered are trade offs between payload, compression bitrate, distortion induced by the insertion of the hidden data or the watermark and robustness of watermarked images in the watermarking context. We have proposed several joint JPEG2000 compression and data hiding schemes. Two of these joint schemes are watermarking systems. All the embedding strategies proposed in this work are based on Trellis Coded Quantization (TCQ). We exploit the channel coding properties of TCQ to reliably embed data during the quantization stage of the JPEG2000 part 2 codec.

Orientadores: Yuzo Iano, Ana Lúcia Mendes Cruz Silvestre da Silva / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-18T18:35:20Z (GMT). No. of bitstreams: 1 Leite_RicardoBarroso_M.pdf: 4189561 bytes, checksum: 31cbbc85c9fa0ec77e54c0fdfac29e3f (MD5) Previous issue date: 2011 / Resumo: Os avanços em telecomunicações e o desenvolvimento dos equipamentos digitais impulsionaram diversas áreas de pesquisa relacionadas à codificação e compressão de imagens. Dentre as áreas de atuação destacam-se as aplicações para dispositivos móveis (celulares, smartphones, iPhones, iPads entre outros), que se caracterizam por baixas taxas de transmissão de dados. Entretanto, imagens codificadas com os padrões atualmente em estado-da-arte apresentam artefatos visuais característicos, como efeito de bloco e ringing. Para contornar a inabilidade das transformadas ortogonais em lidar com a geometria, é proposto na literatura o uso de dicionários wedgelet e da decomposição cartoon-textura. Nesse contexto, propõe-se um método de codificação híbrido wedgelet-wavelet inédito que preserva componentes de cartoon e textura, superando em qualidade visual ao uso de dicionários isolados e se aproximando do desempenho de sistemas de codificação completos, tais como o padrão JPEG 2000. Os ganhos de desempenho, principalmente em qualidade visual das imagens reconstruídas para baixas taxas de dados, indicam que a metodologia apresentada pode vir a ser incluída em sistemas de transmissão com restrições de largura de banda, como por exemplo a TV digital móvel / Abstract: Advances in telecommunications and the development of digital equipments have improved several research areas related to coding and image compression. Among these application fields are the mobile devices (cellphones, smartphones, iPhones, iPad, and others), characterized by low data transmission rates. However, images encoded by state-of-the-art standards present characteristic visual artifacts, like blocking and ringing effects. To surpass the disadvantages of orthogonal transforms in dealing with geometry, wedgelets dictionaries and cartoon-texture decomposition are proposed in literature. In this context, a new hybrid wedgelet-wavelet coding method that preserves cartoon and texture components is proposed, achieving better visual quality than the use of isolated dictionaries, approaching the performance of complete codification systems, such as the JPEG 2000. The performance gains, especially concerning visual quality of the reconstructed images using low data rates, show that this methodology might be adopted in restricted bandwidth transmission systems, such as the digital mobile TV / Mestrado / Telecomunicações e Telemática / Mestre em Engenharia Elétrica

The purpose of this thesis is to present a modified version of the algorithm used in the open source differencing tool zdelta, entitled "iDelta". This algorithm will manage file data and will be built specifically to difference images in the Photoshop file format.

Compressed images are used very frequently in interactive applications in digital video broadcasting. New methods increasing efficiency of the image transmission in digital video broadcasting networks are proposed. Adaptive spatial filtering methods have been proposed for enhancement of the visual perception of the compressed images. New optimalization method is based on application of the filtering algorithms on more compressed images (data size are reduced). Visual quality enhancement is processed in interactive application. Further, new compression methods JPEG2000 and H.264 for image compression have been analysed. Novel compound image compression method for standard and high spatial television resolution is proposed in the thesis.

The aim of this master's thesis was to design, develop and test a method for lossless image compression. The theoretical part includes a description of selected exiting methods such as RLE, MTF, adaptive arithmetic coding, color models used in LOCO-I and JPEG 2000, predictors MED, GAP and laplacian pyramid. The conclusion includes a comparison of various combinations of chosen approaches and overall efficiency compared with PNG and JPEG-LS.

This dissertation develops several new algorithms to solve existing problems in practical application of the previously developed PCA+JPEG2000, which has shown superior rate-distortion performance in hyperspectral image compression. In addition, a new scheme is proposed to facilitate multi-temporal hyperspectral image compression. Specifically, the uniqueness in each algorithm is described as follows. 1. An empirical piecewise linear equation is proposed to estimate the optimal number of major principal components (PCs) used in SubPCA+JPEG2000 for AVIRIS data. Sensor-specific equations are presented with excellent fitting performance for AVIRIS, HYDICE, and HyMap data. As a conclusion, a general guideline is provided for finding sensor-specific piecewise linear equations. 2. An anomaly-removal-based hyperspectral image compression algorithm is proposed. It preserves anomalous pixels in a lossless manner, and yields the same or even improved rate-distortion performance. It is particularly useful to SubPCA+JPEG2000 when compressing data with anomalies that may reside in minor PCs. 3. A segmented PCA-based PCA+JPEG2000 compression algorithm is developed, which spectrally partitions an image based on its spectral correlation coefficients. This compression scheme greatly improves the rate-distortion performance of PCA+JPEG2000 when the spatial size of the data is relatively smaller than its spectral size, especially at low bitrates. A sensor-specific partition method is also developed for fast processing with suboptimal performance. 4. A joint multi-temporal image compression scheme is proposed. The algorithm preserves change information in a lossless fashion during the compression. It can yield perfect change detection with slightly degraded rate-distortion performance.