Catálogo de publicaciones - libros

This paper presents a novel technique of image Enhancement which can be widely used in medical and biological imaging to improve the image quality. The principle objective of enhancement is to process an image so that the result is more suitable than the original image for a specific application. Image enhancement enhances weak edges or weak features in an image while keeping strong edges or features. All existing methods of image enhancement decompose images in a separable fashion, and thus cannot use the geometric information in the transform domain to distinguish weak edges from noises. Therefore, they either amplify noises or introduce visible artifacts, when they are applied to noisy images. The NonSubsampled Contourlet transform built upon NonSubsampled pyramids and NonSubsampled directional filter banks can provide a shift invariant directional multi resolution image representation. The geometric information is gathered pixel by pixel from the NonSubsampled Contourlet Transform coefficients. The proposed method achieved better enhancement results than the wavelet based methods of enhancement.

Image matching is an inarguably important operation for many practical sophisticated systems in machine vision and medical diagnosis. Many gray-level image matching applications use the sum-of-squared-difference (SSD) or sum-of-absolute-differences (SAD) , which are very sensitive to noise. Almost all images have some kind of noise, which causes the matching tasks significantly difficulty. In this paper we explore a new, less noise sensitive image-matching technique. It uses non linear similarity measure min or median on interest points to find a match. The algorithm has been tested using a range of images with different gaussian noise. The result shows a significant improvement over traditional Euclidean distance measure technique for image matching.

Multi-Document summarization strictly needs distinguishing the similarity between sentences & paragraphs of texts because repeated sentences shouldn’t exist in final summary so in order to applying this anti-redundancy we need a mechanism that can determining semantic similarities between sentences and expressions and paragraphs and finally between texts. In this paper it’s used a fuzzy approach to determining this semantic similarity. We use fuzzy similarity and fuzzy approximation relation for gaining this goal. At first, lemma of Persian words and verbs obtained and then synonyms create a fuzzy similarity relation and via that relation the sentences with near meaning calculated with help of fuzzy proximity relation. So we can produce an anti-redundant final summary that have more valuable information.

The modeling of diagnostic systems of taxonomies using fuzzy logic is presented in this paper. Specifically the taxonomies system solo is studied, which that can be applied in a wide range of fields of diagnostic science. The intelligent system that is developed based on the presented modeling can make easier the use of diagnostic systems in education since the test correction is extremely hard and demands experts that are not always available. Additionally, the rate of the extraction of results is a reason for using and distributing such tools (diagnostic systems) in the educational process. It is very useful for e-learning systems [ 1 ], [ 2 ], and distance diagnostics systems.

This paper describes the development of a memory-based lemmatiser for Afrikaans called Lia . The paper commences with a brief overview of Afrikaans lemmatisation and it is indicated that lemmatisation is seen as a simplified process of morphological analysis within the context of this paper. This overview is followed by an introduction to memory-based learning — the machine learning technique that is used in the development of the Afrikaans lemmatiser. The deployment of Lia is then discussed with specific emphasis on the format of the training and testing data that is used. The Afrikaans lemmatiser is then evaluated and it is indicated that Lia achieves a linguistic accuracy figure of over 90%. The paper concludes with some ideas on future work that can be done to improve the linguistic accuracy of the Afrikaans lemmatiser.

Arabic is a Semitic language that is rich in its morphology. Arabic has very numerous and complex morphological rules. Arabic morphological analysis has gained the focus of Arabic natural language processing research for a long time in order to achieve the automated understanding of Arabic. With the recent technological advances, Arabic natural language generation has received attentions in order to allow for a room for wider applications such as machine translation. For machine translation systems that support a large number of languages, interlingua-based machine translation approaches are particularly attractive. In this paper, we report our attempt at developing a rule-based Arabic morphological generator for task-oriented interlingua-based spoken dialogues. Examples of morphological generation results from the Arabic morphological generator will be given and will illustrate how the system works. Nevertheless, we will discuss the issues related to the morphological generation of Arabic words from an interlingua representation, and present how we have handled them.

This paper describes design and creation of a multilingual parallel corpus for South African languages. One of the applications of the corpus, namely, the induction of a part-of-speech tagger for Afrikaans from the data, is presented in the paper. Development of the Afrikaans part-of-speech tagger is based on a modified method for induction of linguistic tools from parallel corpora originally proposed by Yarowsky and Ngai (2001).

In music search and recommendation methods used in the present time, a general filtering method that obtains a result by inquiring music information and recommends a music list using users’ profiles is used. However, this filtering method presents a certain difficulty to obtain users’ information according to their circumstances because it only considers users’ static information, such as personal information. In order to solve this problem, this paper defines a type of context information used in music recommendations and develops a new filtering method based on statistics by applying it to a content-based filtering method. In addition, a recommendation system using a content-based filtering method that was implemented by a ubiquitous computing technology was used to support service mobility and distribution processes. Based on the results of the performance evaluation of the system used in this study, it significantly increases not only the satisfaction for the music selection, but also the quality of services.

Undersampling and aliasing occurs frequently in many imaging systems leading to degradation of image quality. Super-resolution attempts to reconstruct a high-resolution image by fusing the incomplete scene information contained in the sequence of under-sampled images. This paper investigates iterative approaches to super-resolution. We propose algorithm that utilises a relatively small number of low-resolution images and is computationally inexpensive. Experimental results of reconstruction are presented.

Human computer interaction is concerned in the way Users (humans) interact with the computers. Some users can interact with the computer using the traditional methods of a keyboard and mouse as the main input devices and the monitor as the main output device. Due to one or another reason, some users are enable to interact with machines using a mouse and keyboard device, hence there is need for special devices. If we use computer for more time it is really difficult to sit on the chair, keeping hands continuously on the keyboard or mouse and keep watching the monitor. To relax our body and interact comfortably with computer, we need some special device or method, so that computer will understand and accept commands without keyboard or by clicking mouse. Speech Recognition System helps users who are unable to use traditional Input and Output (I/O) devices. Since four decades, man has been dreaming for an “intelligent machine” which can master the natural speech. In its simplest form, this machine should consist of two subsystems, namely Automatic Speech Recognition (ASR) and Speech Understanding (SU). The goal of ASR is to transcribe natural speech while SU is to understand the meaning of the transcription. Recognising and understanding a spoken sentence is obviously a knowledge-intensive process, which must take into account all variable information about the speech communication process, from acoustics to semantics and pragmatics