CODE AND CODING SYSTEMS

Author: Fariña Martínez Antonio.
Year: 2004.
University: A CORUÑA [www.udc.es].
Place of defense: Facultad de Informática.
Place of preparation: Facultad de Informática.

Summary: The textual databases are growing in recent years due to the proliferation of biliotecas digital databases documentaries, and especially by the continued strong growth that the site is maintained. The comresión emerges as an ideal solution that can reduce storage space and operations of E / S, with a consequent benefit for the transmission of information via a network. While compression born in the first part of the twentieth century in the last décda appear new compression techniques based on Huffman, using palabrad with symbols to compress. These new techniques not only improve the ability of other compression methods well known (eg Ziv-Lempel), but also the ability to perform searches within the compressed text, without decompress, so much faster than when such searches being made on the text. Continuing with the idea of compression based on words, in this theory developed four new compression techniques that provide start a new family of compressors based on the use of dense codes. Of these four techniques, two are semiestáticas and two are dynamic. Their names are: End-Tagged Dense Code (s c) -Dense Code, Dynamic End-Tagged Dense and Dynamic Code (s c) -Dense Code. In addition it has also been developed for the first time a compressor dynamic byte oriented and based on words, it uses as Huffman encoding scheme. The experimental results obtained by comparing our body against real compressors have shown that these pose a significant contribution in the field of compression for both systems-oriented Text Retrieval, as in systems for data transmission, as our compressors compress more and more efficiently than many of the compressors currently in use (gzip, compress, etc.).

Author: FERRÁNDEZ AGULLÓ FRANCISCO.
Year: 2004.
University: ALICANTE [www.ua.es].
Place of defense: UNIVERSIDAD DE ALICANTE.
Place of preparation: ESCUELA POLITÉCNICA SUPERIOR.

Summary: The increase in the speed of computers and new algorithms attack that arise continually must make continued increase in the security of cryptographic systems. What is safe today may not be Buddha after a few years. This need to increase security is not without its drawbacks. The complexity achieved today (Order exponential) systems elíptica curve does not seem improved because there are certain attacks that are always applicable groups (of the square root, as Pollard-r). Another option to increase security is to increase the size of systems with larger keys. But this requires the use of large keys if you use the DLP (Discrete Logarithm Problem), which is based DSA (Digital Signature Algorithm) or ellFP (Integer Factorization Problem), which is based on RSA. The ECDLP (Elliptic Curve DLP) keys used much shorter but requires large computational resources as it is necessary to calculate the order of the group of points of the curve elíptica sound. With the objective of designing systems safer without difficulties Reviewed, three unidirectional functions that increase the size of the problem with key short and almost no resource requirements. All three systems are designed as a complex mathematical problems can define functions with unidirectional trap necessary for the design of public-key cryptographic systems. The first proposal is a cryptographic system of linear arrays based on elliptic curve which consists of an application of arrays built so polin6mica and n-tuplas points in n-tuplas points. This feature gets enhance the security of the cryptographic schemes by increasing levels necessary to resolve it. Also shown adaptable to the most important protocols based on ECDLP. The second proposal lies in a cryptographic system nonlinear elliptic curve and matrices consisting of the formal exponentiation elements constructed with two scalar and a point of a curve elíptica. This system is equivalent to ECDLP but, in any case, requires greater effort by computer iteration while, through rapid exponentiation algorithms, implementation of the feature is efficient. The third proposal is a cryptographic system nonlinear elliptic curve and matrices formal triangular block that is based on exponentiation array formal constructed with blocks of scalar and a block of points in a Elliptic Curve. This system improves safety as much as is needed almost no resources. In addition, it shows the applicability of cryptographic protocols through the exchange of keys and encryption. In addition to the description and analysis of the three proposed systems, there is a practical implementation of these protocols and cryptographic adapted, along with a number of considerations of implementation.

Author: ÁLVAREZ SÁNCHEZ RAFAEL IGNACIO.
Year: 2005.
University: ALICANTE [www.ua.es].
Place of defense: ESCUELA POLITÉCNICA SUPERIOR.
Place of preparation: ESCUELA POLITÉCNICA SUPERIOR.

Summary: This work is motivated primarily by the excellent properties of cryptographic and randomness observed in certain constructions upper triangular matrices block elements in Zp, where p cousin. This technique matrix is interesting not only because of its potential crypto, but also because of its high level of flexibility by allowing the resulting algorithm is adaptable to various types of platforms (hardware, software, low-cost devices, etc.). Already diverse operational requirements (high performance, high security, low memory use, etc.). Moreover, not only has the potential to create pseudo generators or ciphers in flux, there are mechanisms to create other types of primitives from the same technical basis. This offers great advantages in terms of integrating different cryptographic service in a single component. It has proposed a generator pseudoaleatroria with desirable characteristics to be used as a generator in a sequence cifrante cryptosystem encryption flow additive binary. It is based on the powers of the upper triangular matrices block defindas in Zp. As they take these powers, you get a sequence of matrices great period and very good properties in terms of aleatroriedad. To achieve yields comparable to cryptosystem encryption flow RC4, has designed an optimized implementation in Z2 based on the concept of arrays packaged and operations between words binary bits that also allows for the direct use of the registers large in the recent architectures. There has also been a hardware implementation of optimization in Z2, the potential for exploiting parallelism present in the platform. The results have been very good: the statistical characteristics exceeds in most cases the cryptosystems reference, while the performance is quite comparable to cryptosystems as efficient as RC4.

Author: OLIVER GIL JOSE SALVADOR.
Year: 2005.
University: POLITÉCNICA DE VALENCIA [www.upv.es].
Place of defense: UNIVERSIDAD POLITÉCNICA DE VALENCIA.
Place of preparation: UNIVERSIDAD POLITÉCNICA DE VALENCIA.

Author: FERNÁNDEZ CÓRDOBA CRISTINA.
Year: 2005.
University: AUTÓNOMA DE BARCELONA [www.uab.es].
Place of defense: ETSE.
Place of preparation: ESCUELA DE POSTGRADO.

Summary: The year 1972, the spacecraft Mariner 9 transmitted images of Mars using the code Reed-Muller of order 1 RM (1.5). These codes are of particular interest because of the ease of the construction process, coding and decoding. From the year 1994, it opened a new door to a investigaicón the theory codes. It was proved that some very important non-linear codes (Preaparta, Kerdock, etc.) had code structure Z-4-lineales. From that moment on, it also began to study the relationship of codes Reed-Muller with quaternary codes. At appeared a couple of families of codes quaternary related Reed-Muller, codes (QRM (r, m) and the codes ZRM (r, m). In this thesis we study in depth these two families. Regards codes QRM (r, m), we will describe a generalization, build a class where, for each value rym, any C code that belongs wing class that meets C Release 2 is exactly the code RM (R m). Generalizaremos basic properties of codes QRM (r, m) at the codes of the class. was demonstrated that all codes Preparata-like i Kerdock-like are the picture via the application of Gray codes in the classroom. were also calculated the rank i dimension kernel code of this class. Eventually, we may find different constructions of these codes and the creation of nested chains and properties of these channels related to the duality and the minimum distance of the codes that comprise it. in the literature, there are two different definitions of codes ZRM (r, m). We will serve this notation ZRM (r, m), for codes defined by Hammons, Kumar Caldermakr, Sloane Sole and the year 1994 i ZRM * (r, m) for codes subsequently defined by Zhe-Xian Wan year 1997. We have demonstrated that these codes match if and only if r = 0,1,2, m, m +1 are exactly the values for which the image of these codes via Gray is the implementation of a code Reed-Muller. studied properties of these two families. images of Gray implementation of the codes ZRM (r, m) with linear codes and we estimate its size and the range and size of the nucleus the images of codes ZRM * (r, m).

Author: QUEIRUGA DIOS M. ARACELI.
Year: 2005.
University: SALAMANCA [www.usal.es].
Place of defense: FACULTAD DE CIENCIAS.
Place of preparation: FACULTAD DE CIENCIAS.

Summary: In this paper we study the so-called public-key cryptosystems' backpack 'high density. In particular, it performs a detailed analysis of cryptosystem of Chor-Rivest. These kinds of cryptosystems are called backpack for being based on the computational problem of the same name. The first one was proposed by Hellman and Merkle and subsequently broken by Shamir and Brickell. Later, in 1985, Rivest Chor and other proposed cryptosystem that has remained invulnerable until Vaudenay, in 2001, proposed a cryptanalysis for some of the original parameters. In more precise, and Chor Rivest cryptosystem proposed a kind of backpack, based on the arithmetic of finite bodies, F_ (q) ^ h, q being a cousin, or the power of a cousin, close to 200 h = q a minor entire near 25. This gives cryptosystem for the generation of the keys is required in the computation of logarithms in the body finite. Problem which, as is known, is considered very difficult from the computational point of view. However, the safety of this cryptosystem is not based on the problem of computing discrete logarithms, but the difficulty of solving a problem backpack high density. The attack proposed by Vaudenay determines a private key corresponding to that used by the recipient. Once the standard is calculated, determining the permutation used in the phase of generating passwords. Knowing both parameters, Vaudenay demonstrates how it is possible to determine a key corresponding to the private key original. Then he examines the ca so where the finite parameters qyh body on which it is carried out by the arithmetic of cryptosystem of Chor-Rivest are not to allow such conduct, in an effective way, the attack Vaudenay and to continue being efficient processes for encryption and decryption. Such parameters are q = 409 h = 17. The memory is accompanied by the procedures to help set up an implementation practice Maplae of different aspects studied.

Author: RUIZ HIDALGO JAVIER.
Year: 2006.
University: POLITÉCNICA DE CATALUÑA [www.upc.edu].
Place of defense: Aula de Teleen.-·Edif. B-3-Campus Nord.
Place of preparation: D5 108 Nord.

Summary: This paper studies the utility to exploit the synergy between the representations compression and indexing video sequences. The study was separated into two main tasks. In the first task, the performances of compression have been analyzed to produce representations of indexing more optimized. In the second task, the performances of indexing has been exploited to generate representations compression more efficient. In this thesis, the term ârepresentación of compresiónâ refers to all those data, normally called bitstream, which describes the digital content of a compact, using the fewest possible bits and allowing turn functionality viewing content. Moreover, the term ârepresentación of indexaciónâ refers to all those data structures, descriptors or extra information has been extracted from the digital content to provide indexing features, summary, search or acquisition. In recent years, both performances has been studied independently, and therefore representations compression and indexing have been derived or generated using different algorithms and schemes. However, both representations describing the same digital content and therefore it is normal to think that each of them can benefit from the other. The first task of the theory, representations of compression are analyzed to create representations of flat video more robust and optimized. In particular, the two representations are investigating the mosaics to create more efficient representations. The encoding scheme by mosaics of standard MPEG-4 is extended to use the mosaic, not only in coding, but to summarize and describe levels of video. The representation of indexing is proposed in this paper is based on the content of the scene as it is analyzed to separate objects of foreground and background. The mosaic is used to represent objects in the background while other representations are called key regions, are created to represent objects in the foreground. In addition, this paper mosaics are improved by neighboring operators so that they can better represent the background of the scene without entailing any loss in its coding efficiency. The mosaics are also used in the algorithm for extracting key regions, allowing the segmentation and description of objects in the foreground. The representation of indexing proposed in this thesis represents a compact summary of the plane video analyzed. Moreover, the representation of indexing information can be enriched by adding movement which can have a first idea of how the objects in the foreground are moving in the scene of the video plane. In the second task, looks at improving the efficiency of compression performances, in particular the standard H.264, using representations of indexing. She explores four different techniques to prove that the representations indexing, even if they are drawn to cover other features can be exploited to increase the efficiency of current video coders. The first technical proposal, the different transitions are encoded video using the extra information of a descriptor transitions. The second technical reformulates part of the estimated movement of a hybrid encoder as a classic problem of search and acquisition. This second technique improves the selection of frames of reference in the long term through representations indexing low-level descriptors such as color. .