Imagens de documentos são documentos digitalizados com conteúdo textual. Estes documentos são compostos de caracteres e diagramação, apresentando características comuns entre si, como a presença de bordas e limites no formato de cada caractere. A relação entre as características das imagens de documentos e os impactos do processo de compressão com respeito à fidelidade visual são analisadas nesse trabalho. Métricas objetivas são empregadas na análise das características das imagens de documentos, como a medida da atividade da imagem (IAM) no domínio espacial dos pixels, e a verificação da medida de atividade espectral (SAM) no domínio espectral. Os desempenhos das técnicas de compressão de imagens baseada na transformada discreta de cosseno (DCT) e na transformada discreta de Wavelet (DWT) são avaliados sobre as imagens de documentos ao aplicar diferentes níveis de compressão sobre as mesmas, para cada técnica. Os experimentos são realizados sobre imagens digitais de documentos impressos e manuscritos de livros e periódicos, explorando texto escritos entre os séculos 16 ao século 19. Este material foi coletado na biblioteca Brasiliana Digital (www.brasiliana.usp.br), no Brasil. Resultados experimentais apontam que as medidas de atividade nos domínios espacial e espectral influenciam diretamente a fidelidade visual das imagens comprimidas para ambas as técnicas baseadas em DCT e DWT. Para uma taxa de compressão fixa de uma imagem comprimida em ambas técnicas, a presença de valores superiores de IAM e níveis menores de SAM na imagem de referência resultam em menor fidelidade visual, após a compressão. / Document images are digitized documents with textual content. These documents are composed of characters and their layout, with common characteristics among them, such as the presence of borders and boundaries in the shape of each character. The relationship between the characteristics of document images and the impact of the compression process with respect to visual fidelity are analyzed herein. Objective metrics are employed to analyze the characteristics of document images, such as the Image Activity Measure (IAM) in the spatial domain, and assessment of Spectral Activity Measure (SAM) in the spectral domain. The performance of image compression techniques based on Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) are evaluated from document images by applying different compression levels for each technique to these images. The experiments are performed on digital images of printed documents and manuscripts of books and magazines, exploring texts written from the 16th to the 19th century. This material was collected in the Brasiliana Digital Library in Brazil. Experimental results show that the activity measures in spatial and spectral domains directly influence the visual fidelity of compressed images for both the techniques based on DCT and DWT. For a fixed compression ratio for both techniques on a compressed image, higher values of IAM and low levels of SAM in the reference image result in less visual fidelity after compression.

Este trabalho tem por objetivo comparar os seguintes métodos de compressão de imagens de impressão digital: transformada discreta do cosseno (DCT), transformada de wavelets de Haar, transformada de wavelets de Daubechies e transformada de wavelets de quantização escalar (WSQ). O propósito da comparação é identificar o método que resulta numa menor perda de dados, para a maior taxa de compressão possível. São utilizadas as seguintes métricas para avaliação da qualidade da imagem para os métodos: erro quadrático médio (ERMS), a relação sinal e ruído (SNR) e a relação sinal ruído de pico (PSNR). Para as métricas utilizadas a DCT apresentou os melhores resultados, seguida pela WSQ. No entanto, o melhor tempo de compressão e a melhor qualidade das imagens recuperadas avaliadas pelo software GrFinger 4.2, foram obtidos com a técnica WSQ. / This research aims to compare the following fingerprint image compression methods: the discrete cosseno transform (DCT), Haar wavelet transform, Daubechies wavelets transform and wavelet scalar quantization (WSQ). The main interest is to find out the technique with the smallest distortion and higher compression ratio. Image quality is measured using peak signal-to-noise ratio (PSNR), signal-to-noise ratio (SNR) and root mean square (ERMS). Image quality using these metrics showed best results for the DCT followed by WSQ, although the WSQ had the best compression time and presented the best quality when evaluated by the GrFinger 4.2 software.

Imagens de documentos são documentos digitalizados com conteúdo textual. Estes documentos são compostos de caracteres e diagramação, apresentando características comuns entre si, como a presença de bordas e limites no formato de cada caractere. A relação entre as características das imagens de documentos e os impactos do processo de compressão com respeito à fidelidade visual são analisadas nesse trabalho. Métricas objetivas são empregadas na análise das características das imagens de documentos, como a medida da atividade da imagem (IAM) no domínio espacial dos pixels, e a verificação da medida de atividade espectral (SAM) no domínio espectral. Os desempenhos das técnicas de compressão de imagens baseada na transformada discreta de cosseno (DCT) e na transformada discreta de Wavelet (DWT) são avaliados sobre as imagens de documentos ao aplicar diferentes níveis de compressão sobre as mesmas, para cada técnica. Os experimentos são realizados sobre imagens digitais de documentos impressos e manuscritos de livros e periódicos, explorando texto escritos entre os séculos 16 ao século 19. Este material foi coletado na biblioteca Brasiliana Digital (www.brasiliana.usp.br), no Brasil. Resultados experimentais apontam que as medidas de atividade nos domínios espacial e espectral influenciam diretamente a fidelidade visual das imagens comprimidas para ambas as técnicas baseadas em DCT e DWT. Para uma taxa de compressão fixa de uma imagem comprimida em ambas técnicas, a presença de valores superiores de IAM e níveis menores de SAM na imagem de referência resultam em menor fidelidade visual, após a compressão. / Document images are digitized documents with textual content. These documents are composed of characters and their layout, with common characteristics among them, such as the presence of borders and boundaries in the shape of each character. The relationship between the characteristics of document images and the impact of the compression process with respect to visual fidelity are analyzed herein. Objective metrics are employed to analyze the characteristics of document images, such as the Image Activity Measure (IAM) in the spatial domain, and assessment of Spectral Activity Measure (SAM) in the spectral domain. The performance of image compression techniques based on Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) are evaluated from document images by applying different compression levels for each technique to these images. The experiments are performed on digital images of printed documents and manuscripts of books and magazines, exploring texts written from the 16th to the 19th century. This material was collected in the Brasiliana Digital Library in Brazil. Experimental results show that the activity measures in spatial and spectral domains directly influence the visual fidelity of compressed images for both the techniques based on DCT and DWT. For a fixed compression ratio for both techniques on a compressed image, higher values of IAM and low levels of SAM in the reference image result in less visual fidelity after compression.

This thesis presents new approaches to image registration and super-resolution mosaicing as well as their applications. Firstly, a feature-based image registration method is proposed for a multisensor surveillance system that consists of an optical camera and an infrared camera. By integrating a non-rigid object tracking technique into this method, a novel approach to simultaneous object tracking and multisensor image registration is proposed. Based on the registration and fusion of multisensor information, automatic face detection is greatly improved. Secondly, some extensions of a gradient-based image registration method, called inverse compositional algorithm, are proposed. These extensions include cumulative multi-image registration and the incorporation of illumination change and lens distortion correction. They are incorporated into the framework of the original algorithm in a consistent manner and efficiency can still be achieved for multi-image registration with illumination and lens distortion correction. Thirdly, new super-resolution mosaicing algorithms are proposed for multiple uncompressed and compressed images. Considering the process of image formation, observation models are introduced to describe the relationship between the superresolution mosaic image and the uncompressed and compressed low-resolution images. To improve the performance of super-resolution mosaicing, a wavelet-based image interpolation technique and an approach to adaptive determination of the regularization parameter are presented. For compressed images, a spatial-domain algorithm and a transform-domain algorithm are proposed. All the proposed superresolution mosaicing algorithms are robust against outliers. They can produce superresolution mosaics and reconstructed super-resolution images with improved subjective quality. Finally, new techniques for super-resolution sprite generation and super-resolution sprite coding are proposed. Considering both short-term and long-term motion influences, an object-based image registration method is proposed for handling long image sequences. In order to remove the influence of outliers, a robust technique for super-resolution sprite generation is presented. This technique produces sprite images and reconstructed super-resolution images with high visual quality. Moreover, it provides better reconstructed low-resolution images compared with low-resolution sprite generation techniques. Due to the advantages of the super-resolution sprite, a super-resolution sprite coding technique is also proposed. It achieves high coding efficiency especially at a low bit-rate and produces both decoded low-resolution and super-resolution images with improved subjective quality. Throughout this work, the performance of all the proposed algorithms is evaluated using both synthetic and real image sequences.

Image coding plays a key role in multimedia signal processing and communications. JPEG2000 is the latest image coding standard, it uses the EBCOT (Embedded Block Coding with Optimal Truncation) algorithm. The EBCOT exhibits excellent compression performance, but with high complexity. The need to reduce this complexity but maintain similar performance to EBCOT has inspired a significant amount of research activity in the image coding community. Within the development of image compression techniques based on wavelet transforms, the EZW (Embedded Zerotree Wavelet) and the SPIHT (Set Partitioning in Hierarchical Trees) have played an important role. The EZW algorithm was the first breakthrough in wavelet based image coding. The SPIHT algorithm achieves similar performance to EBCOT, but with fewer features. The other very important algorithm is SBHP (Sub-band Block Hierarchical Partitioning), which attracted significant investigation during the JPEG2000 development process. In this thesis, the history of the development of wavelet transform is reviewed, and a discussion is presented on the implementation issues for wavelet transforms. The above mentioned four main coding methods for image compression using wavelet transforms are studied in detail. More importantly the factors that affect coding efficiency are identified. The main contribution of this research is the introduction of a new low-complexity coding algorithm for image compression based on wavelet transforms. The algorithm is based on block dividing coding (BDC) with an optimised packet assembly. Our extensive simulation results show that the proposed algorithm outperforms JPEG2000 in lossless coding, even though it still leaves a narrow gap in lossy coding situations

This thesis relates to research and development in the field of 3D mesh data for computer graphics. A review of existing storage and manipulation techniques for mesh data is given followed by a framework for mesh editing. The proposed framework combines complex mesh editing techniques, automatic level of detail generation and mesh compression for storage. These methods work coherently due to the underlying data structure. The problem of storing and manipulating data for 3D models is a highly researched field. Models are usually represented by sparse mesh data which consists of vertex position information, the connectivity information to generate faces from those vertices, surface normal data and texture coordinate information. This sparse data is sent to the graphics hardware for rendering but must be manipulated on the CPU. The proposed framework is based upon geometry images and is designed to store and manipulate the mesh data entirely on the graphics hardware. By utilizing the highly parallel nature of current graphics hardware and new hardware features, new levels of interactivity with large meshes can be gained. Automatic level of detail rendering can be used to allow models upwards of 2 million polygons to be manipulated in real time while viewing a lower level of detail. Through the use of pixels shaders the high detail is preserved in the surface normals while geometric detail is reduced. A compression scheme is then introduced which utilizes the regular structure of the geometry image to compress the floating point data. A number of existing compression schemes are compared as well as custom bit packing. This is a TIF funded project which is partnered with Unlimited Realities, a Palmerston North software development company. The project was to design a system to create, manipulate and store 3D meshes in a compressed and easy to manipulate manner. The goal is to create the underlying technologies to allow for a 3D modelling system to become integrated into the Umajin engine, not to create a user interface/stand alone modelling program. The Umajin engine is a 3D engine created by Unlimited Realities which has a strong focus on multimedia. More information on the Umajin engine can be found at www.umajin.com. In this project we propose a method which gives the user the ability to model with the high level of detail found in packages aimed at creating offline renders but create models which are designed for real time rendering.

Made available in DSpace on 2014-12-17T14:08:51Z (GMT). No. of bitstreams: 1 DanielTS_DISSERT.pdf: 2906320 bytes, checksum: 0844f1ce3a3b057671d92efe62faa050 (MD5) Previous issue date: 2012-04-06 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / In this work, spoke about the importance of image compression for the industry, it is known that processing and image storage is always a challenge in petrobr?s to optimize the storage time and store a maximum number of images and data. We present an interactive system for processing and storing images in the wavelet domain and an interface for digital image processing. The proposal is based on the Peano function and wavelet transform in 1D. The storage system aims to optimize the computational space, both for storage and for transmission of images. Being necessary to the application of the Peano function to linearize the images and the 1D wavelet transform to decompose it. These applications allow you to extract relevant information for the storage of an image with a lower computational cost and with a very small margin of error when comparing the images, original and processed, ie, there is little loss of quality when applying the processing system presented . The results obtained from the information extracted from the images are displayed in a graphical interface. It is through the graphical user interface that the user uses the files to view and analyze the results of the programs directly on the computer screen without the worry of dealing with the source code. The graphical user interface, programs for image processing via Peano Function and Wavelet Transform 1D, were developed in Java language, allowing a direct exchange of information between them and the user / Neste trabalho, apresenta-se a import?ncia da compress?o de imagens para a ind?stria de petr?leo, sabe-se que o processamento e armazenamento de imagens ? sempre um desafio nas grandes empresas de petr?leo, para otimizar o tempo de armazenamento e armazenar um n?mero m?ximo de imagens e dados. ? exposto algumas ferramentas para o processamento e armazenamento de imagens no dom?nio wavelet. A proposta apresentada baseia-se na Fun??o de Peano e na transformada wavelet 1D. O sistema de armazenamento tem como objetivo a otimiza??o do espa?o computacional, tanto para o armazenamento como para transmiss?o de imagens. Sendo necess?rio para isso, a aplica??o da Fun??o de Peano para linearizar as imagens com m?xima concentra??o de pontos vizinhos e a transformada wavelet 1D para decomp?-la. Estas aplica??es permitem extrair informa??es relevantes para o armazenamento de uma imagem com um menor custo computacional e com uma margem de erro muito pequena quando se compara as imagens, original e processada, ou seja, h? pouca perda de qualidade ao aplicar o sistema de processamento apresentado. Os resultados obtidos a partir das informa??es extra?das das imagens s?o apresentados numa interface gr?fica. ? atrav?s da interface gr?fica que o usu?rio visualiza as imagens e analisa os resultados do programa diretamente na tela do computador sem a preocupa??o de lidar com os c?digos fontes. A interface gr?fica, os programas de processamento de imagens via Fun??o de Peano e a TransformadaWavelet 1D foram desenvolvidos em linguagem java, possibilitando uma troca direta de informa??es entre eles e o usu?rio

Orientador: Yuzo Iano / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-16T03:46:01Z (GMT). No. of bitstreams: 1 Araujo_AndreFilgueirasde_M.pdf: 3583920 bytes, checksum: afbfc9cf561651fe74a6a3d075474fc8 (MD5) Previous issue date: 2010 / Resumo: Este trabalho tem como objetivo principal a elaboração de um novo algoritmo responsável por tornar mais eficiente a estimação de movimento do codec Dirac. A estimação de movimento é uma etapa crítica à codificação de vídeo, na qual se encontra a maior parte do seu processamento. O codec Dirac, recentemente lançado, tem como base técnicas diferentes das habitualmente utilizadas nos codecs mais comuns (como os da linha MPEG). O Dirac objetiva alcançar eficiência comparável aos melhores codecs da atualidade (notadamente o H.264/AVC). Desta forma, este trabalho apresenta inicialmente estudos comparativos visando à avaliação de métodos de estado da arte de estimação de movimento e do codec Dirac, estudos que fornecem a base de conhecimento para o algoritmo que é proposto na sequência. A proposta consiste no algoritmo Modified Hierarchical Enhanced Adaptive Rood Pattern Search (MHEARPS). Este apresenta desempenho superior aos outros algoritmos de relevância em todos os casos analisados, provendo em média complexidade 79% menor mantendo a qualidade de reconstrução. / Abstract: The main purpose of this work is to design a new algorithm which enhance motion estimation in Dirac video codec. Motion estimation is a critical stage in video coding, in which most of the processing lies. Dirac codec, recently released, is based on techniques different from the usually employed (as in MPEG-based codecs). Dirac video codec aims at achieving efficiency comparable to the best codecs (such as H.264/AVC). This work initially presents comparative studies of state-of-the-art motion estimation techniques and Dirac codec which support the conception of the algorithm which is proposed in the sequel. The proposal consists in the algorithm Modified Hierarchical Enhaced Adaptive Rood Pattern Search (MHEARPS). This presents superior performance when compared to other relevant algorithms in every analysed case, providing on average 79% less computations with similar video reconstruction quality. / Mestrado / Telecomunicações e Telemática / Mestre em Engenharia Elétrica

Orientadores: Yuzo Iano, Rangel Arthur / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-21T22:39:06Z (GMT). No. of bitstreams: 1 Beltrao_GabrielTedgue_M.pdf: 7074938 bytes, checksum: 0a2d464733f2fb5dcc14430cc1844758 (MD5) Previous issue date: 2012 / Resumo: As transformadas derivadas da DCT são amplamente utilizadas para compressão de vídeo. Recentemente, muitos autores têm destacado que os resíduos de predição normalmente apresentam estruturas direcionais que não podem ser eficientemente representadas pela DCT convencional. Nesse contexto, muitas transformadas direcionais têm sido propostas como forma de suplantar a deficiência da DCT em lidar com tais estruturas. Apesar do desempenho superior das transformadas direcionais sobre a DCT convencional, para a sua aplicação na compressão de vídeo é necessário avaliar o aumento no tempo de codificação e a complexidade para sua implementação. Este trabalho propõe um rápido algoritmo para se estimar as direções existentes em um bloco antes da aplicação das transformadas direcionais. O codificador identifica as direções predominantes em cada bloco e aplica apenas a transformada referente àquela direção. O algoritmo pode ser usado em conjunto com qualquer proposta de transformadas direcionais que utilize a técnica de otimização por taxa-distorção (RDO) para a seleção da direção a ser explorada, reduzindo a complexidade de implementação a níveis similares a quando apenas a DCT convencional é utilizada / Abstract: DCT-based transforms are widely adopted for video compression. Recently, many authors have highlighted that prediction residuals usually have directional structures that cannot be efficiently represented by conventional DCT. In this context, many directional transforms have been proposed as a way to overcome DCT's deficiency in dealing with such structures. Although directional transforms have superior performance over the conventional DCT, for application in video compression it is necessary to evaluate increase in coding time and complexity for its implementation. This work proposes a fast algorithm for estimating blocks directions before applying directional transforms. The encoder identifies predominant directions in each block, and only applies the transform referent to that direction. The algorithm can be used in conjunction with any proposed algorithm for directional transforms that uses the rate-distortion optimization (RDO) process for selection of the direction to be explored; reducing implementation complexity to similar levels when only conventional DCT is used / Mestrado / Telecomunicações e Telemática / Mestre em Engenharia Elétrica

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.