Browsing by Author "De Silva, CR"

Now showing 1 - 20 of 24

item: Conference-Abstract
Architecture of ensemble neural networks for risk analysis
De Silva, END; Ranasinghe, KAMK; De Silva, CR; Thurairajah, N
Assembling of neural networks referred to as “Ensemble neural networks” consist with many small “expert networks” that learn small parts of the complex problem, which are established by decomposing it into its sub levels. Ensemble neural network architecture has been proposed to solve complex problems with large numbers of variables. In this paper, this architecture is used to analyze maintainability risks of high-rise buildings. An ensemble neural network that consists with four expert networks to represent four building elements namely roof, façade, basement and internal areas is developed to forecast the maintenance efficiency (ME) of buildings. The model is tested and the results showed good performance. The model is further validated using a real case study.
item: Conference-Full-text
Ballotx - automated ballot paper counting system
(Computer Science & Engineering Society c/o Department of Computer Science and Engineering, University of Moratuwa., 2010-09) Azhar, MNM; Rathnayake, DRMGM; Madusanka, AIV; Rajapakshe, L; De Silva, CR; Ransiri, S; Gunasekara, C; Wijegunawardana, P; Pavalanathan, U
Even though the technology has evolved dramatically throughout the last decade or so, many countries have stuck with the traditional ballot paper counting process. With BallotX we propose a viable alternative to reduce time and resources consumed in the traditional manual process by introducing image processing into this domain. This research paper discusses the ways in which the above can be achieved for varying ballot papers and elections. It also addresses the issues of transparency and reliability which has top priority when transforming such sensitive manual process to an automated one
item: Article-Abstract
Cloud basis function neural network:Amodified RBF network architecture for holistic facial expression recognition
De Silva, CR; Ranganath, S; De Silva, LC
The paper presents novel modifications to radial basis functions (RBFs) and a neural network based classifier for holistic recognition of the six universal facial expressions from static images. The new basis functions, called cloud basis functions (CBFs) use a different feature weighting, derived to emphasize features relevant to class discrimination. Further, these basis functions are designed to have multiple boundary segments, rather than a single boundary as for RBFs. These new enhancements to the basis functions along with a suitable training algorithm allow the neural network to better learn the specific properties of the problem domain. The proposed classifiers have demonstrated superior performance compared to conventional RBF neural networks as well as several other types of holistic techniques used in conjunction with RBF neural networks. The CBF neural network based classifier yielded an accuracy of 96.1%, compared to 86.6%, the best accuracy obtained from all other conventional RBF neural network based classification schemes tested using the same database.
item: Conference-Full-text
Computer assisted railway control system for Sri Lanka railways
(Computer Science & Engineering Society c/o Department of Computer Science and Engineering, University of Moratuwa., 2010-09) Sayooran, N; Fernando, JS; Mohamed, MA; Thambugala, LB; De Silva, CR; Gunasekara, C; Wijegunawardana, P; Pavalanathan, U
This paper describes about the design and implementation of the Computer Assisted Railway Control System for Sri Lanka Railways. Currently, the train traffic control system of Sri Lanka Railways has two control units. They are Outstation and Metropolitan control units. In the outstation unit, controllers spend more time in manually plotting the TDG (Time Distance Graph) than decision making and for reporting purposes the same data are re-entered. In the metropolitan unit, switch boards in illumination based panel is used to monitor and control train traffic. This panel has many drawbacks such as congested layout, unclear information and unreliable bulbs and switches. So at the end of the day, all these problems contribute to trains getting delayed. The proposed system is built using 4-tier networked architecture which addresses the aforementioned drawbacks and bottlenecks. Features such as TDG, TDG Designer, Track Designer, Dispatcher Control View, Rules Editor, Train Timetable, Traffic Control Playback, Auto Train Supervision and Reporting have helped to improve the overall operation of Centralized Train Traffic Controlling of Sri Lanka Railways. Moreover, the solution comprises a high performance and more reliable hardware unit with higher I/O capacity that provides a hardware interface between the proposed system and the existing electromechanical interlocking system. The solution was successfully tested at Maradana Control Office and Dcmatagoda Workshop with the presence of senior officials from Sri Lanka Railways. The test results obtained depict that the system has helped to improve the overall train controlling operations of Sri Lanka Railways and proved its potential for implementation.
item: Conference-Full-text
Coyote - computer vision based speed camera
(Computer Science & Engineering Society c/o Department of Computer Science and Engineering, University of Moratuwa., 2010-09) Rathnayaka, ADMP; Induwara, BAS; Udayangana, MPS; Samarakoon, TS; De Silva, CR; Gunasekara, C; Wijegunawardana, P; Pavalanathan, U
This paper describes about a new way of vehicle speed detection which uses computer vision and image processing techniques. The paper proposes how we could use a real time video stream and calculate speeds of vehicles that are present in the video, by overcoming issues like partial and full occlusions. Also this paper contains descriptions of each module of the proposed system and their functionalities.
item: Conference-Full-text
District cooling systems for mega projects in Sri Lanka
(Department of Architecture, University of Moratuwa, Sri Lanka, 2017-10-20) Kankanamge, S; De Silva, N; De Silva, CR; Wijesundara, J; Dayaratne, R
In marching towards sustainable development after 30 years of civil war, Sri Lanka is currently undergoing a boom time with regard to its construction industry. A number of mega projects are ongoing in the capital city of the country. Thus, managing the high operational costs of these upcoming mega facilities will become a challenging task in near future. Air conditioning is identified as a prominent building facility in terms of high operational costs. The aim of the research is to study the effect of District Cooling Systems (DCS) on the reduction of energy demand arising from the operation of mega facilities and, thus, subsequently reducing the operational costs. A qualitative approach was selected as the research methodology to achieve the research objectives. An expert survey was carried out to identify the dominant enablers of DCS implementation in the country. The expert survey included professionals from multi-disciplinary backgrounds such as investors, HVAC contractors, DCS engineers and facility managers. Furthermore, an in-depth case study was carried out with the intention of validating the potential energy saving with the proposed DCS approach. Six existing large-scale buildings which are located close to each other were selected for the case study. The results showed that the implementation of DCS in Sri Lanka (as an energy conservational approach) has tremendous potential. They showed a 25% reduction in energy demand when maintaining a temperature of 170C. Thus, DCS can be proposed as an effective approach to reducing the operational costs of mega facilities. Furthermore, the adaption of these advanced technologies in infrastructure developments will enhance the attraction of potential international investors to the country.
item: Conference-Extended-Abstract
An extensible computer vision application for blood cell recognition and analysis
(2007) De Silva, CR; Priyanka, GPM; Seneviratne, OW; Silva, RKOH; Soysa, WVD
Fast and cost-effective production of blood count reports is of paramount importance in the healthcare industry especially during pandemic situations. The traditional method of manual count under the microscope yield inaccurate results and put an intolerable amount of stress on the Medical Laboratory Technicians (MLT). On the other hand, although there are hardware solutions such as the "Automated Haematology Counter", developing countries like Sri Lanka are not capable of deploying such prohibitively expensive machines in every hospital and laboratory in the country. As a solution to this problem, this research project aims to provide a software-based cost-effective and an efficient alternative in recognising and analysing blood cells.
item: Conference-Extended-Abstract
Face recognition using informative vector machines
(2007) Wimalawarne, KADNK; De Silva, CR
In the recent past kernel methods have been successfully applied to face recognition. We present a novel approach in frontal face recognition with informative vector machine, a sparse Gaussian process kernel classifier. Informative vector machine has the ability to provide more sparse solutions than the widely used kernel classifiers like support vector machines.
item: Conference-Full-text
A faster image registration and stitching algorithm
(Computer Science & Engineering Society c/o Department of Computer Science and Engineering, University of Moratuwa., 2009-07) Senarathne, N; Ransiri, S; Balasooriya, A; Arangala, P; De Silva, CR; Nanayakkara, V; Gunasinghe, UP
This paper presents a new algorithm for Image Registration and Stitching. The algorithm is designed to be extremely efficient and fast in its execution and is intended for use in stitching images extracted from a video stream of a camera. This algorithm is not universally applicable to all the image registration and stitching problems. It is customized to be used to generate single images of surfaces such as a conveyor belt or undercarriage of vehicles, which cannot be captured by a single photo. The algorithm works by extracting edges of the two images to be registered. Then it selects a reference section from the first image and search in the second image where it finds the best match for that section. The best match is the least difference score. The section in the second image which gives the least difference score is then identified as the registration point between the two images. The speed of the algorithm is improved by using heuristics to restrict the search space in the second image. We present full details of how the extraction of the heuristic is done from the inputs and how it drastically reduces the execution time of the algorithm. The implementation of the algorithm was done in MATLAB and the details of the MATLAB functions are listed in the paper. The paper also contains a full section on comparing our algorithm with a set of existing algorithms. Our algorithm outperforms the existing ones for all the common image sizes.
item: Conference-Extended-Abstract
Image retrieval and indexing system for data mining
(2007) Fernando, PBC; Gunasekera, GSM; Gunasinghe, SR; Udawatta, L; De Silva, CR
Given the limitations of human perceptions it would be impossible to classify the images manually. The IRIS (Image Retrieval and Indexing System for Data Mining) system that we have developed is a solution! to this problem. "QBIC" (Query by Image Content) system of IBM, "Content-Based Visual Query" Project of the University of Columbia as well as "CLUE" and "SIMPLIcity" of Pennsylvania State University are of considerable academic value.
item: Thesis-Full-text
Impact of leadership styles on employee performance :
(2018) Karunajeewa, D; De Silva, CR
The topic of the research is impact of leadership styles on employee performance at ABC Company. The main reasons as to why the researcher was guided to conduct a study based on the above problem was low performance of employees, job dissatisfaction among employees, increasing absenteeism and turnover rate and considerable increase of client complaints. As the objectives of this study the below statements can be stated. To understand the current styles of leadership within the organization, to identify the factors affecting employee performance, to identify the most determinant factor for employee performance and to recommend an effective leadership style which can enhance the employee performance of ABC Company. The sample size of the research was 80 employees where the population was 80 employees which is same as the sample size. Researcher has used two types of data collection methods which are primary and secondary. Furthermore, as the data analysis method the researcher has used qualitative and quantitative methods to analyze the collected data. Finally, the data presentation has done by using linear scales, tables, charts, pie graphs and scatter graphs. The key findings of the research are as follows. As qualitative findings it proved that the independent variables which are democratic, laissez faire and transformational leadership styles have a positive relationship with the dependent variable which is employee performance and also autocratic and transactional leadership styles have negative relationship with employee performance. As the qualitative analysis also, the findings were proved.
item: Conference-Extended-Abstract
Intellitraffic computer vision based hightwa y traffic parameter estimation system
(2007) De Silva, CR; Dharmawardana, KATS; Ratnayake, ADSE; Rodrigo, SAKC; Sandeepani, LAA
The Traffic Control System of a country is a main area that has the potential to affect the development of a country in a very extensive scale. The existing traffic control systems in Sri Lanka, UniRoadl and UniRoad2 are highly dependent on the statistical traffic information gathered from time to time using manual methods(using people etc.) which are more difficult to rely on. The main drawback of all these systems is that they cannot adapt signal timing dynamically according to the situation. A system with the capability of altering signal timing as per real time situations will be more effective and productive. For that it needs a proper traffic parameter estimation system which should offer features like reliability and efficiency to a far beyond the manual methods. This will save time wasted at highway while reducing fuel consumption, noise and air pollution in urban areas. Various parties like Traffic Police, the Road Development Authority and ultimately the whole community will be benefited from such a system. The basis of such a system would be a method for reliably approximating the traffic flow in real time, at a low cost. This research paper discusses how to approximate the traffic flow from video streams and why some of the methods attempted are not suitable for such a system.
item: Conference-Abstract
Maintainability approach for lean maintenance
(2016-06-21) De Silva, N; Ranasinghe, KAMK; De Silva, CR
The concept of lean maintenance is promoted to eliminate maintenance waste and to maximize efficiency of the manufacturing process for overall cost reduction. The optimum usage of labour,material and technology is thus, proposed to eliminate waste during the maintenance process. This paper proposes a maintainability approach to minimize maintenance waste in order to promote lean maintenance concept to the construction industry. The approach has widened the focus of existing lean maintenance concept by moving it boundaries from the maintenance phase to early phases of the development process. Eight challenges of lean maintenance have been established from fifty one maintainability causes of high-rise buildings. Their impact to maintenance cost was evaluated using a case study and found 40% reduction in the maintenance cost as wastage
item: Conference-Full-text
Maintainability approach for lean maintenance
(2012-06) De Silva, N; Ranasinghe, M; De Silva, CR
The concept of lean maintenance is promoted to eliminate maintenance waste and to maximise efficiency of the manufacturing process for overall cost reduction. The optimum usage of labour, material and technology is thus, proposed to eliminate waste during the maintenance process. This paper proposes a maintainability approach to minimise maintenance waste in order to promote lean maintenance concept to the construction industry. The approach has widened the focus of existing lean maintenance concept by moving it boundaries from the maintenance phase to early phases of the development process. Eight challenges of lean maintenance have been established from fifty one maintainability causes of high-rise buildings. Their impact to maintenance cost was evaluated using a case study and found 40% reduction in the maintenance cost as wastage.
item: Article-Abstract
Maintainability of condominiums constructed for low-income families in Sri Lanka
De Silva, END; Sampath, DBD; De Silva, CR
Sri Lanka now has a prime market for “vertical growth” and the private sector has become the main developer. However, the government has also contributed to these developments through the condominiums constructed for low income families. These were built for the purpose of re-settling people from shanties and recently, Tsunami victims. These condominiums have shown significant failures and problems. This paper looks at some of these problems; specially maintenance. Three types of condominiums were selected to study. Interviews were conducted with management corporations, selected occupants and professionals in the relevant authorities. The study focused on understanding maintainability problems related to technical facilities, management, social condition, and legal background. Twenty five technical management, social and legal related problems were observed that occur during the maintenance and operational stages. Some of them have become a huge burden to the government. Further, the study identified eight strategies from substantive experts to eliminate these issues now and in future attempts.
item: Conference-Extended-Abstract
A mutipurpose optimized directshow filter for videomixing applications
(2007) Udawatte, L; De Silva, CR; Chathuranga, SAK; Danansuriya, LNP; Ganegedara, KMHS; Gunasinghe, UP
A video mixer [1] is used to select one or many video streams from different video sources and mix them together to form a single video output. The currently available video mixers use specific hardware devices of very high cost. Our solution "Chamelean" is a cost effective real-time video mixing and broadcasting software which can run on a regular personal computer using generic devices. Chameleon performs the mixing of multiple video streams such as applying transitions from one video to another, alpha blending of multiple video streams and compositing videos within each other as well as applying text, effects and images to videos.
item: Conference-Full-text
A novel method and tool for detection of tumors and tissue abnormalities MRIplus for MRI image classification and segmentation
(IEEE, 2016-05) Jayamal, MD; Kurukulasooriya, KD; Weerasiri, SRBVT; De Silva, CR; Jayasekara, AGBP; Bandara, HMND; Amarasinghe, YWR
A single MRI scan generates a large number of images of various cross sections of the body. This large set of accumulated data makes manual analysis time consuming thus a smart tool for screening is vital. This paper presents a novel classification and segmentation method which has the ability to identify white matter in MRI mages. Based on those findings, a supporting web based tool for the MRI image analysis and screening is introduced.
item: Conference-Extended-Abstract
Outstation railway supervision system for Sri Lanka railways
(2010) Sayooran, N; Fernando, JS; Mohamed, MA; Thambugala, LB; De Silva, CR
This paper discusses the design of a centralized control and data analysis system for railway traffic of Sri Lnaka Railways. The overall solution comprises of a central database server system with redundancy, Controller's Controlling Interface for remote monitoring and controlling and a TCP/IP network which interconnects all the above mentioned components of the total systems. Outstation Railway Control Unit is a manually operated system with a Frequency Division Multiplexing (FDM) voice communication system integrated to a computer system that runs on an Internet Protocol (IP) based network. The man drawback of the existing system is lack of accuracy and permits less time for the controllers to make decisions in controlling the train. Since the given standards graph paper is not design accurately, as the divisions for the hours along the time axis are not uniformly distributed. Therefore the predictions done by the controller are not perfect. All the reports are generated by pen and paper manually which have made the work of the controller more complicated. At the end of the day all these drawbacks results in trains getting delayed. The proposed solution automates the train controlling processes, saving more time for decision making of the controllers and minimizing train getting delayed.
item: Article-Full-text
Risk analysis in maintainability of high-rise buildings under tropical conditions using ensemble neural network
(Emerald Group Publishing Limited, 2016) De Silva, N; Ranasinghe, M; De Silva, CR
Purpose – The aim of this research study is to develop a risk-based framework that can quantify maintainability to forecast future maintainability of a building at early stages as a decision tool to minimize increase of maintenance cost. Design/methodology/approach – A survey-based approach was used to explore the risk factors in the domain of maintainability risks under tropical environmental conditions. The research derived ten risk factors based on 58 identified causes related to maintainability issues as common to high-rise buildings in tropical conditions. Impact of these risk factors was evaluated using an indicator referred to as the “maintenance score (MS)” which was derived from the “whole-life maintenance cost” involved in maintaining the expected “performance” level of the building. Further, an ensemble neural network (ENN) model was developed to model theMSfor evaluating maintainability risks in high-rise buildings. Findings – Results showed that predictions from the model were highly compatible and in the same order when compared with calculations based on actual past data. It further showed that, maintainability of buildings could be improved if the building was designed, constructed and managed properly by controlling their maintainability risks. Originality/value – The ENN model was used to analyze maintainability of a high-rise building. Thus, it provides a useful tool for designers, clients, facilities managers/maintenance managers and users to analyze maintainability risks of buildings at early stages.
item: Article-Abstract
Risk factors affecting building maintenance under tropical conditions
(2016-06-21) De Silva, N; Ranasinghe, KAMK; De Silva, CR
Purpose – The factors contributing to maintenance cost include the life-long environmental factors that the building is exposed to during its useful life as well as parameters relating to the design, construction and maintenance processes. The performance of risk associated with these factors decides the future maintenance requirements of buildings. Ignoring these risks factors generally result in high maintenance costs. The purpose of this paper is to identify the maintainability risk factors and their implications on maintenance costs. Design/methodology/approach – A survey based methodology was used to establish the risks factors of building maintainability. A sample of high-rise buildings was selected to evaluate these risk factors that were identified through an extensive literature review and survey of substantive experts. Further, these risks factors were applied to a real case study. Findings – A total of ten risk factors of maintainability were identified in relation to high-rise buildings. They were analyzed using a case study and the implications on maintenance cost were evaluated. The results show that there is a considerable reduction in maintenance cost when these risks are minimized. Research limitations/implications – The research was limited to the USA and its six biggest cities. The results presented are predominantly qualitative and the results rely on extensive evaluation. Originality/value – The paper provides useful information to the designers and users on the maintenance problems related to buildings. Also it highlights critical risk conditions that should be considered to minimize maintenance costs.