An expert system for structure damage assessment

An expert system called Sperill-II is introduced for the damage assessment of existing structures using the knowledge of experienced structural engineers. Fuzzy sets and Dempster and Shafer's theory are used in this inexact inference method.     

An image-based automatic Arabic translation system

In this paper, we present a system that automatically translates Arabic text embedded in images into English. The system consists of three components: text detection from images, character recognition, and machine translation. We formulate the text detection as a binary classification problem and apply gradient boosting tree (GBT), support vector machine (SVM), and location-based prior knowledge to improve the F1 score of text detection from 78.95% to 87.05%. The detected text images are processed by off-the-shelf optical character recognition (OCR) software. We employ an error correction model to post-process the noisy OCR output, and apply a bigram language model to reduce word segmentation errors. The translation module is tailored with compact data structure for hand-held devices. The experimental results show substantial improvements in both word recognition accuracy and translation quality. For instance, in the experiment of Arabic transparent font, the BLEU score increases from 18.70 to 33.47 with use of the error correction module.     

An Image-Based Virtual Reality Prototype System

The most important goal of virtual reality is to create a virtual world computers where users are allowed to view the environment and control the virtual objects interactively.Traditionally,virtual reality systems use 3D computer graphics to model and render a virtual environment in real time.However,this approach usually requires laborious modeling and expensive special-purpose rendering hardware.Image-based rendering is a new approach in composing a virtual environment in which a set of panoramic images is used to compose the virtual environment and walking in the space is accomplished by “hopping”to different panoramic points.This paper introduces an experimental image-based VR system.The techniques utilized in the system,in particular the authoring and interactive control tools of the system,are described in detail.     

A robust approach for seismic damage assessment

We present a robust approach based on a convex analysis to take into account uncertainties appearing in evaluations of damage to structures during an earthquake. Two kinds of uncertainties are introduced: on the excitation and on the structure. The results obtained by the convex analysis are compared with a Monte Carlo simulation to evaluate the conservatism of the method. The analysis is applied to a reinforced concrete shear wall which is modeled as a single degree of freedom system. The non-linearity of the structure is taken into account by an uncertainty on a linearized system. An extension to multi-degrees of freedom systems is presented to allow the treatment of more complex structures, with an application to portal frames. The results of this analysis consist in an evaluation of the maximum possible damage of the structure according to the uncertainties taken into account.     

A robust hybrid image-based modeling system

This paper presents a new robust image-based modeling system for creating high-quality 3D models of complex objects from a sequence of unconstrained photographs. The images can be acquired by a video camera or hand-held digital camera without the need of camera calibration. In contrast to previous methods, we integrate correspondence-based and silhouette-based approaches, which significantly enhances the reconstruction of objects with few visual features (e.g., uni-colored objects) and improves surface smoothness. Our solution uses a mesh segmentation and charting approach in order to create a low-distortion mesh parameterization suitable for objects of arbitrary genus. A high-quality texture is produced by first parameterizing the reconstructed objects using a segmentation and charting approach, projecting suitable sections of input images onto the model, and combining them using a graph-cut technique. Holes in the texture due to surface patches without projecting input images are filled using a novel exemplar-based inpainting method which exploits appearance space attributes to improve patch search, and blends patches using Poisson-guided interpolation. We analyzed the effect of different algorithm parameters, and compared our system with a laser scanning-based reconstruction and existing commercial systems. Our results indicate that our system is robust, superior to other image-based modeling techniques, and can achieve a reconstruction quality visually not discernible from that of a laser scanner.     

An automatic parking system using an optimized image-based fuzzy controller by genetic algorithms

An automatic parking system of a car-like mobile robot is an important issue in commercial applications. An image-based fuzzy controller for an automatic parking system of a car-like mobile robot was developed in previous work, where the membership functions were tuned by experimentally. The aim of this paper is to optimize the parameters of the membership functions, which were performed in previous work, using a genetic algorithm against the complicated tuning of the controller. The details of GA implementation, such as the design parameters and choice of fitness function, are described. Simulation results illustrate the effectiveness of the developed schemes.     

An ontology of slums for image-based classification

Information about rapidly changing slum areas may support the development of appropriate interventions by concerned authorities. Often, however, traditional data collection methods lack information on the spatial distribution of slum-dwellers. Remote sensing based methods could be used for a rapid inventory of the location and physical composition of slums. (Semi-)automatic detection of slums in image data is challenging, owing to the high variability in appearance and definitions across different contexts. This paper develops an ontological framework to conceptualize slums using input from 50 domain-experts covering 16 different countries. This generic slum ontology (GSO) comprises concepts identified at three levels that refer to the morphology of the built environment: the environs level, the settlement level and the object level. It serves as a comprehensive basis for image-based classification of slums, in particular, using object-oriented image analysis (OOA) techniques. This is demonstrated by with an example of local adaptation of GSO and OOA parameterization for a study area in Kisumu, Kenya. At the object level, building and road characteristics are major components of the ontology. At the settlement level, texture measures can be potentially used to represent the contrast between planned and unplanned settlements. At the environs level, factors which extend beyond the site itself are important indicators, e.g. hazards due to floods plains and marshy conditions. The GSO provides a comprehensive framework that includes all potentially relevant indicators that can be used for image-based slum identification. These characteristics may be different for other study areas, but show the applicability of the developed framework.     

An assessment of the Fe-Mn system

Thermochemical and phase equilibrium data for the Fe-Mn system has been fitted to the subregular solution model. The model parameters indicate that all the phases have a negative deviation from ideal solution behaviour but the effect is very weak for the bcc, fcc and liquid phases.     

Object manipulation system based on image-based reinforcement learning

Intelligent Service Robotics - Advances in reinforcement learning algorithms allow robots to learn complex tasks such as object manipulation. However, most of these tasks have been implemented only...     

A transportable system of models for natural resource damage assessment

A computer model system has been developed for assessment of natural resource economic damages resulting from spills of oil and hazardous materials in marine and freshwater environments. The model can address a wide range of spatial and temporal scales. The equations describing the motion of both pollutants and biota are solved in three dimensions. The model can simulate continuous releases of a contaminant, with representation of complex coastal boundaries, variable bathymetry, multiple shoreline types, and spatially variable ecosystem habitats. A graphic user interface (GUI) gives the facile control of the system, as well as the ability to display elements of the underlying geographical information system (GIS) data base. The system has been developed for lakes, rivers, estuarine and marine environments. The structure of the system is such that transport to new geographic regions requires only the development of the appropriate physical, toxicological, biological, and economic data sets.     

基于参考图像的彩墨画快速渲染法

通过对彩墨画创作过程的模拟,提出一种基于参考图像的彩墨画快速渲染算法,其模拟过程分为铺底模拟和细节刻画两个阶段。在铺底阶段,利用平均曲率流与冲击滤波的组合对图像进行细节简化,并通过对颜色空间的截取和拉伸数字模拟彩墨画的留白特征,同时提供基于PatchMatch算法的颜色学习着色方法;在细节刻画阶段,针对墨水的扩散现象提出一个随机模拟扩散的公式,并给出模拟彩墨画描边手法的边界处理方法,最后进行纸张效果合成。实验结果表明,该算法能够快速地将参考图像渲染成具有彩墨画风格的艺术作品。     

基于红外图像的车底藏人检测系统设计

针对以往检测车底藏人依靠人工查看获取到的车底画面效率不高和工作量大的问题,通过组合滤波、改进背景差分、车底线面特征提取、ROI划分等方法,设计出一套基于红外图像车底藏人自动检测系统。实验表明：该检测系统具有准确率高、适用性广、实用性高等优点,能够实现目标的自动检测。     

An environmental assessment system for environmental technologies

A new model for the environmental assessment of environmental technologies, EASETECH, has been developed. The primary aim of EASETECH is to perform life-cycle assessment (LCA) of complex systems handling heterogeneous material flows. The objectives of this paper are to describe the EASETECH framework and the calculation structure. The main novelties compared to other LCA software are as follows. First, the focus is put on material flow modelling, as each flow is characterised as a mix of material fractions with different properties and flow compositions are computed as a basis for the LCA calculations. Second, the tool has been designed to allow for the easy set-up of scenarios by using a toolbox, the processes within which can handle heterogeneous material flows in different ways and have different emission calculations. Finally, tools for uncertainty analysis are provided, enabling the user to parameterise systems fully and propagate probability distributions through Monte Carlo analysis.     

An object-oriented system for ergonomic risk assessment

Abstract: During the last decade, cumulative trauma disorders have become a major concern both to occupational health and safety professionals and manufacturing industries. This paper will present the development of a knowledge-based system for predicting the potential of developing strain injuries among employees exposed to repetitive and forceful activities. This potential is identified through the use of an expert system incorporating a fuzzy set algorithm. Recommendations for the prevention of the resulting CTDs are then provided by the system. The system executes on a microcomputer platform, and was constructed utilising a commercial object-oriented expert system shell.     

Online learning design of an image-based facial expression recognition system

In order to serve people and support them in daily life, a domestic or service robot needs to accommodate itself to various individuals. Emotional and intelligent human–robot interaction plays an important role for a robot to gain attention of its users. Facial expression recognition is a key factor in interactive robotic applications. In this paper, an image-based facial expression recognition system that adapts online to a new face is proposed. The main idea of the proposed learning algorithm is to adjust parameters of the support vector machine (SVM) hyperplane for learning facial expressions of a new face. After mapping the input space to Gaussian-kernel space, support vector pursuit learning (SVPL) is employed to retrain the hyperplane in the new feature space. To expedite the retraining step, we propose to retrain a new SVM classifier by using only samples classified incorrectly in previous iteration in combination with critical historical sets. After adjusting the hyperplane parameters, the new classifier will recognize more effectively previous unrecognizable facial datasets. Experiments of using an embedded imaging system show that the proposed system recognizes new facial datasets with a recognition rate of 92.7%. Furthermore, it also maintains a satisfactory recognition rate of 82.6% of old facial samples.