Abstract

It is well known that the performance of solar cells may significantly suffer from local electric defects. Accordingly, infrared thermography (i.p. lock-in thermography) has been intensely applied to identify such defects as hot spots. As an imaging method, this is a fast way of module characterization. However, imaging leads to a huge amount of data, which needs to be investigated. An automatized image analysis would be a very beneficial tool but has not been suggested so far for lock-in thermography images. In this manuscript, we describe such an automatized analysis of solar cells. We first established a robust algorithm for segmentation (or recognition) for both, the PV-module and the defects (hot spots). With this information, we then calculated a parameter from the IR-images, which could be well correlated with the maximal power (Pmpp) of the modules. The proposed automatized method serves as a very useful foundation for faster and more thorough analyses of IR-images and stimulates the further development of quality control on solar modules.

Abstract

The design of an oral corpus and the processes of registering, codifying and treating the materials in order to build a useful resource for linguistic analysis prompt numerous decisions regarding theory and methodology. This article is focused on those stages of corpus construction which are more clearly conditioned by the computational processing necessary to make it functional.  In order to adequately match the initial expectations and the real possibilities of using the tool, each feature we intend to codify must be measured against the workload and the means required to do so. Therefore, it is essential to take into account the available possibilities of processing and exploitation as they have a crucial impact on decisions regarding the corpus’ construction. Based on experience acquired in the construction of the ESLORA corpus, the present article looks into some of the problems arising in the process of designing an oral corpus, such as the delicacy with which oral phenomena are represented, the segmentation of the discourse, the coexistence of different simultaneous tagging systems and the particularities of annotation in a bilingual or multilingual context.

Abstract

The generation of anatomical models is one the most important concern to biomedical researchers as well as to medical doctors, due to needed to understand the human tissues. Is know that the soft tissues like heart, brain, prostate and hard tissues like jaw, bones, skull, etc are structures of complex morphologies, so, the anatomical models generation is not an easy and trivial task. Currently, this task has benefited of advances of imaging diagnostic, which permit obtain cross and longitudinal sections of human body. In this research, we describe a method to obtain 3D discrete models of human body given by a dataset of medical images. Five main modules were implemented in prototype software: (1) Reading and 3D reconstruction of Computerized Axial Tomography and Magnetic Resonance Images. (2) Preprocessing techniques for improve the low medical images quality by using enhancement algorithms to reduce image noise and to increase structures contrast. (3) Combined segmentation techniques for tissue identification, which were applied through a multi-stage approach. (4) Post processing techniques to improve segmented volumes and (5) Exportation task of volumes to readable formats by Computer Aided Design (CAD) tools to be later analyzed by numerical methods. The performance of our method is shown on several medical examples and the techniques were validated using statistical descriptors to compare our models with models from free databases. Results showed that the implemented techniques generate precise and useful models for numerical analysis and medical survey, planning and surgery in a short processing time.

Abstract

Technological progress has profoundly changed the way personal data are collected, accessed and used. Those data make possible unprecedented customization of advertising which, in turn, is the business model adopted by many of the most successful Internet companies. Yet measuring the value being generated is still a complex task. This paper presents a review of the literature on this subject. It has been found that the economic analysis of personal information has been conducted up to now from a qualitative perspective mainly linked to privacy issues. A better understanding of a quantitative approach to this topic is urgently needed.

Abstract

This paper discusses the detection of moving objects (being a crucial part of driver assistance systems) using monocular or stereoscopic computer vision. In both cases, object detection is based on motion analysis of individually tracked image points (optical flow), providing a motion metric which corresponds to the likelihood that the tracked point is moving. Based on this metric, points are segmented into objects by employing a globally optimal graph-cut algorithm. Both approaches are comparatively evaluated using real-world vehicle image sequences. Document type: Part of book or chapter of book

Abstract

Proceedings of: 23rd International Conference on Industrial, Engineering & Other Applications of Applied Intelligent Systems (IEA-AIE 2010) Córdoba-Spain, June 04-06, 2010 This paper presents a new approach for trajectory segmentation in the area of Air Traffic Control, as a basic tool for offline validation with recorded opportunity traffic data. Our approach uses local information to classify each measurement individually, constructing the final segments over these classified samples as the final solution of the process. This local classification is based on a domain transformation using motion models to identify the deviations at a local scale, as an alternative to other global approaches based on combinatorial analysis over the trajectory segmentation domain. This work was supported in part by Projects CICYT TIN2008-06742-C02-02/TSI, CICYT TEC2008-06732-C02-02/TEC, CAM CONTEXTS (S2009/TIC-1485) and DPS2008-07029-C02-02. Publicado Document type: Part of book or chapter of book

Abstract

The use of Computer Vision techniques for the automatic recognition of road signs is fundamental for the development of intelligent vehicles and advanced driver assistance systems. In this paper, we describe a procedure based on color segmentation, Histogram of Oriented Gradients (HOG), and Convolutional Neural Networks (CNN) for detecting and classifying road signs. Detection is speeded up by a preprocessing step to reduce the search space, while classification is carried out by using a Deep Learning technique. A quantitative evaluation of the proposed approach has been conducted on the well-known German Traffic Sign data set and on the novel Data set of Italian Traffic Signs (DITS), which is publicly available and contains challenging sequences captured in adverse weather conditions and in an urban scenario at night-time. Experimental results demonstrate the effectiveness of the proposed approach in terms of both classification accuracy and computational speed. Document type: Part of book or chapter of book