Where Imagination Leads to Reality

Cutting Edge 2009

“Cutting Edge” is an annual exhibition which showcases innovative state-of-the-art IT and business solutions researched and developed at Informatics Institute of Technology (IIT). Cutting Edge 2009 was held on 25th and 26th June 2009 at IIT. Informatics Institute of Technology offers the BSc (Hons) Information Systems with Business Management degree programme in collaboration with University of Westminster, UK. It is a requirement that all students in the final year undertake a research project. Selected final-year projects were displayed at Cutting Edge 2009. These spanned a range of topics under the categories of Surface Computing, Web Services, Multimedia, Mobile Applications, Artificial Intelligence and Embedded Systems. The British Computer Society was invited to judge the best projects that were presented at the exhibition. Virtusa awarded the trophies for Academic Excellence. The following are some of the IT Projects that were presented at Cutting Edge 2009. The President of the British Computer Society, Sri Lankan Chapter, Dr. Gihan Wikramanayake, graced the occasion as the Chief Guest.

iDisplay A Computer – Vision Based System to Retrofit Existing Flat Displays into Interactive Surfaces

iDisplay provides a generic and inexpensive way to retrofit existing regular flat displays into interactive surfaces. The provided solution is simple and the hardware prototype requires only a small number of devices. No expensive sensors or infrared-based techniques were used, only a regular off-the-shelf webcam and a regular LCD monitor. iDisplay allows a regular user to experience novel ways of interacting with a computing system without using any extravagant hardware. Further, it introduces the colour sensing ability for interactive surfaces, which increases the area of applicability to a wide range.

Content-Based Video Editor

Smartcut - the content aware video editor aims to support non-professional video editing by analyzing the content of video clips. This application makes use of the vast amount of computer vision research done in the field to provide a smooth video editing experience to the end user. In a nutshell, this video editor allows users to retouch the content of the video clips by defining objects only in one frame rather than redefining objects in each and every frame. For example, the user can select a vehicle in one frame of a video clip and that will be tracked throughout the whole video. Once the object is tracked, users can apply video effects only to the selected object, such as colour-change, blur.

ChordATune – An Emotion-Based Melody Harmonizer for Piano Music

ChordATune is an interactive accompaniment (harmonizing) tool for piano melodies that generates the most suitable base notes for a given tune according to the emotional factor of the user. This further provides functionalities to create new songs and generate variations of accompaniments (harmonies) according to different styles of music (pop, rock, rumba, swing, guitar fret sheets). ChordATune has been developed using Java, based on AI concepts, machine learning concepts and Hidden Markov Model concepts, along with dynamic programming. Around 250 lead sheets are used to train this system using a data driven and heuristic approach.

Sensor-Based Control Framework for Embedded Systems

The aim of the project is to create a control framework to be used in embedded systems which follows a service-oriented architecture approach. The main objective is to provide an abstraction into the sensors/actuators in a specific device by exposing them as services to the user application. The framework mainly provides a communication architecture where internal device-sensor communication and also inter-device discovery and communication is provided.

SearchDroid – A Discourse Marker-Based Topic Identification and Search Results Refining Assistant

SearchDroid is a software that is targeted to assist researchers by refining search results returned by a search engine to better match their research objectives, based on different topics considered in their research and the relationship(s) between such topics. A linguistics approach called Discourse Parsing has been used to determine the relevance of documents based on topics, their positioning and the type of relationship they share.

The major sponsors for the event were Virtusa Corporation, IFS, Aeturnum and Zone 24X7. We thank our sponsors for their much-valued support which creates an opportunity for young researchers to present their work to industry and academia.