Details of Innovative Projects

Details of Innovative Projects(2013-14)

Project 1
Title of the Project – Finger Vein authentication system using line detection algorithm
Name of the sponsoring company: Apple systems
Period of research: 1 year
Name of the Coordinator: Dr. G. Shanmugasundaram

Abstract

Even though there are many forms of biometrics that are being used in various field of authentication each one has its own drawbacks. On considering all the drawbacks and the requirement for a secured authentication we have proposed an implementation of finger vein authentication in vehicles using line detection algorithm. This authentication standard to be proposed will help the research and developers to improve the overall process of authentication with respect to finger veins .Since there is a rise in heist of many vehicles across the country the introduction of the finger vein authentication system in vehicles provides an high security to all the vehicles. It also helps us to know about the unauthorized person who has accessed the vehicle in the absence of the user. In addition to this new findings may lead to have a set of new research directions in this field.

Project 2
Title of the Project – A Hybrid method for unsupervised object class recognition based on learning object model
Name of the sponsoring company: Apple systems
Period of research: 1 year
Name of the Coordinator: Dr. G. Shanmugasundaram

Abstract

Automatic Detection, Identification and classification of moving live objects is an important part of computer vision applications. The two main goals of object recognition are: image classification and object detection. The detection of shape patterns is an important problem in the field of computer vision because such detection would be robust against variations in the illumination, color and texture of images. Large appearance variations of object instances belonging to the same class which include background clutter, scale and viewpoint variations can render appearances of the same object instance to vary vastly. Many methods pose the shape detection as a task of matching contour shape models in images and they basically utilize hand drawn sketches as reference templates which are said to be semisupervised. For the purpose of rotation invariant and shape invariant object recognition this project aims at the construction of learning object models in an unsupervised way. If training dataset is given the system might learn optimal parameter for the detection of each object class by learning which segments in the object are more discriminative than others, and improve the performance by giving higher weight on more discriminative parts. It is to arrive at a well principle method for unsupervised training for invariant features. High level invariant feature is trained with unlabeled data and a classifier can use these features to classify images.

Project 3
Title of the Project – DABBAWALAPP – on Time Every Time
Name of the sponsoring company: MUMBAIDABBAWALA & TCS
Period of research: 6 Months
Name of the Coordinator: Mr. R. Suresh

Abstract

People of major cities especially Mumbai, work in places which are far from their residences. This means, they must wake up around 4 or 4.30 in the morning. They do not want to pack their lunch prepared too early to their workplace disturbing their loved ones. So there arises the need for “DABBAWALAS” who are monitored by THE DAPPAWALA FOUNDATION (a large network) in MUMBAI. In initial stage, a Dabbawala was communicated and tracked through TELEPHONE facilities (Existing System) which are not up-to-date with the current scenario of development in technology. As per the information obtained from their respective website, they had mentioned that they may use some general and efficient technology for customer satisfaction if needed. New customers may find it hard to seek for a Dappawala near their places. Also, in case of any delay, the customer would not b able to track the Dappawala. For a long period, they had no IT support which is very much needed for them that makes their work lighter.

Project 4
Title of the Project – Cloudinar Halls
Name of the sponsoring company: Unisys, Bangalore
Period of research: 6 Months
Name of the Coordinator: Mr. P.Victer Paul

Abstract

In a large seminar hall, the trainers could not effectively deliver the lecture using a single projector for a huge group of listeners. This problem can be solved by using a projector that can be paired with a tablet or smart phones of the learners, to access projector’s data, using wireless technique. The projector can be integrated with a thin-client to connect with the cloud, where lecture contents are stored, can avoid extra portable storage. The cloud projector can be coupled with a wireless keyboard and mouse could replace the usage of traditional computer based presentations giving presenters the freedom to interact more with the learners.

The Cloudinar halls are basically smart seminar halls which use cloud technology and mainly focuses on solving two problems. One, when the trainer has to address a large crowd the learners at the end away from the projector will not be able to view the projected data clearly. The second issue that this project solves is, whenever a laptop is connected with a projector, so many issues might occur. Connecting a laptop with the projector through a wired medium itself does not seem very easy. Sometimes, the laptop might need any software to project its display. To avoid this, the cloud projectors can directly get the data from the cloud and the trainer can simply present his lecture by accessing the presentation documents from the cloud.

To summarize, the project “Cloudinar Halls” would serve as a solution when there need to conduct a presentations for a huge crowd in an infrastructure-less environment. It is also intended to include many user interactive options between the trainer and the listener devices to make the session more collaborative rather than the single point of lecture.

Project 5
Title of the Project – ARCHAEOPTERYX – A Modern UMS
Name of the sponsoring company: Unisys, Bangalore
Period of research: 1 Year
Name of the Coordinator: Mr. R. Suresh & Mr. P.S. Vijayabaskaran

Abstract

Today, we are living in a peaceful society; It’s all because of the soldiers with gun that matters at the border. Most of their contributions are unknown to the normal citizens of our country. So we have proposed a surveillance vehicle in which we can control its movement and direction by using mobile. This vehicle can monitor armed people at certain distance and inform about them to the soldiers at the border. This model is highly secured and more efficient than any other models. We are glad to say that our idea has been appreciated by the Multi National Company UNISYS.

Product Development