Analysis and Testing

CIBL aims to host research activities directed toward the integration of formal specification methods and analysis with software testing and selective regression testing. Capabilities to support analysis and testing throughout the software lifecycle, from early requirements analysis through operational use are being pursued, as well as extending specification-based testing techniques to be applicable at the level of software architecture.

Led by one of CIBL’s thought leadership, this research makes an attempt to propose a systematic framework or model to facilitate software quality planning for an organization/project. Software Quality Management is referred to as the most challenging activity during the life cycle of a software project. Instead, quality management has not been properly adopted by most of the software development and IT service organizations yet. It all happened due to lack of standard quality management knowledge, and importantly, unavailability of guideline(s) on defining organization specific feasibility criterion towards software quality management activities. This paper proposes a framework or model to provide scientific support to organization/project specific quality planning. The framework makes it easier for real-time software quality managers to obtain defined set of software quality management activities feasible for their respective organization(s), and in addition, it provides them with a scientific methodology of defining the most feasible set of software quality management activities for a specific software project with several organization and project attributes being in serious consideration.

The widespread adoption of Internet technologies and the integration of communication networks into everyday organizational work has led to an increasing interest in the role that information systems and communication technologies can play in supporting collaboration. CRLab aims to perform extensive research work which will take a broad-based approach that focuses as much on the social and organizational factors affecting successful adoption as on the technical challenges for applications and infrastructures. Topics of current interest include the role of technology in supporting distributed and mobile work; the use of virtual meeting technologies in large organizations; infrastructures for group information management; expertise recommendation; virtual worlds supporting working communities; and awareness technologies.

This area focuses on the design, development and evaluation of interactive software systems. We are interested in foundational questions of interaction and usability as well as practical aspects of building effective interactive systems. Future application domains and concerns will include evolutionary software development, expert finding, information visualization, fluid information management, personalized systems, and medical information systems.

Project in Pipeline: Design, Implementation and Performance Analysis of a Multimodal Biometric System

A biometric system may use one or more instances of a single biometric identifier (e.g., multiple impressions of a finger) or it may utilize one or more instances of multiple biometric identifiers (e.g., fingerprint and face images) taken from an individual. Based on the nature of the input, a biometric system can be classified into one of the four categories:

• unibiometric system is a system that uses only a single biometric identifier;
• unimodal biometric system is a subset of a unibiometric system that uses a single instance (snapshot), a single representation, and a single matcher for a recognition decision;
• multibiometric system is a biometric system that uses more than one independent r weakly correlated biometric identifier taken from an individual (e.g. fingerprint and face of the same person, or fingerprints from two different fingers of a person, respectively);
• multimodal biometric system is a superset of a multibiometric system that may use more than one correlated biometric measurement. For example, a multimodal biometric system may be based on multiple ompressions of a finger, multiplr images of a face in video, multiple representations of a single input, multiple matchers of a single representation, or any combination thereof.

Thus, the definition of a unimodal biometric system is the most restrictive and the definition of a multimodal biometric system us the most general. In this research, we will design, implement and analyze the performance of a multimodal biometric system involving fingerprints.

High-security verification applications typically have a very stringent performance requirement (e.g, very low FMR) that a unimodal biometric system is often unable to meet due to limited discriminatory information contained in each biometric. Earlier researches have shown that the information content (number of distinguishable patterns) in two of the most commonly used representations of hand geometry and face are only of the order of 105 and 103, respectively. Therefore, hand geometry and face-based systems are not expected to be sufficiently accurate for high security applications. In addition, although fingerprint and iris possess much larger information content, the available fingerprint-based automatic verification systems are not always able to deal with poor quality images. As a result, they do not meet the high matching accuracy requirements of certain critical applications. The limitations of a unimodal biometric system become more apparent if the system has to operate in the identification mode.

There are, however, a few disadvantages of using a multimodal biometric system. First, a multimodal biometric system is more expensive and requires more computational and storage resources than a unimodal system. Second, when a multimodal system requires that more than one biometric be sensed, it causes some inconvenience to the user and requires additional verification time. For example, in a system that requires both fingerprint and retina images of a person, a user will not only need to touch the fingerprint sensor, but will also need to "peek" into the retinascope. This leads to longer enrollment as well as verification times. Furthermore, an effective fusion scheme is needed to combine the evidence from different modalities. If the multiple modalities are not property combined, the combination may actually degrade system accuracy. But, these disadvantages are almost close to negligible in comparison to its accuracy and advantages, if the multimodal system can be designed and implemented with precision.

Keeping all these pros and cons in mind, CIBL have decided to perform an exclusive research work on multimodal fingerprint based biometric system.