Amit Thawani

Context-aware Timely Information Delivery in Mobile Environments

In mobile environments, transmitting information relevant to an event along with notification of the event has been proven to be an effective means of providing revenue enhancing services. For example, a relevant advertisement can be displayed just before event notification; for instance, a product promotion by Beckham can be shown just before the notification of a goal scored by him. Challenges in achieving real-time search and delivery of information relevant to events as they occur include: (i) predicting the next event so that the appropriate information can be kept ready; (ii) finding information relevant to the context and content of the event; (iii) searching for and bringing the potentially needed information closer to the user location; and (iv) disseminating and displaying relevant information just before the actual event is notified to the user. In this paper, we propose a real-time information delivery system based on an events context and content. The key features of our system are: (i) representing domains with event-based scenarios using statecharts; (ii) using a novel combination of history information and state transitions to predict events; and (iii) real-time delivery of information relevant to an event by prefetching and caching information based on the events predicted to occur in the near future. To illustrate our approach, we focus on the delivery of advertisements relevant to notified events as a specific case of delivery of information relevant to the events. Using an experimental setup, we measured the effectiveness of our approach as compared to a context-unaware approach to event prediction. Experiments demonstrate that our approach has a superior performance, resulting in: (i) better prediction accuracy; (ii) the delivery of the most relevant information most of the time; and (iii) an effective cache management in mobile devices.

Middleware for efficient power management in mobile devices

Mobile computing environment is characterized by limited and dynamically varying available resources and stringent application requirements. Mobile applications are often time-constrained and severely energy constrained as battery power is the only available power source. The power consumed by an application depends on the performance level requested by the user/application. The mobile device can be viewed as the collection of devices such as processor, memory unit, keyboard, display, RF unit etc. Each of these devices can operate at different power states. The power state at which a particular device is operating depends upon the requirement of the application which is using the device. We propose a power management middleware for mobile devices, which not only consider energy savings for the processor but also optimize energy savings for other devices such as display unit, RF unit, keyboard, memory, etc. The proposed approach strives for optimizing the total system power by optimizing the power of individual devices for diverse applications running on the device. Our approach is generic enough to be used on any platform. We have validated our approach by testing two typical mobile applications but our approach can be easily scaled for multiple applications running concurrently.

Web-Based Context Aware Information Retrieval in Contact Centers

One of the important challenges in todays contact center solutions is to increase the speed at which the agents can find information to respond to customer queries. In this paper, we study the issues relevant to information retrieval by contact center agents and propose a solution to address those issues. Our proposed solution is towards: (a) defining user specific, agent specific, and business specific contexts for contact center; (b) using simple mechanisms initially to derive current context of a query and pre-fetch the information and appropriately present the pre-fetched information to the agent; (c) derive contact center specific analyzed context based on past information about customer, agent and business scenarios to understand the context of the query; and (d) using the current and analyzed contexts to retrieve information.

Illumination invariant face recognition for frontal faces using modified census transform

In this paper, a face recognition technique has been proposed by obtaining an illumination invariant representation of the frontal face using the modified census transform and applying the eigen faces approach on the resulting images. An improvement in recognition rate has been observed based on the experiments performed on the frontal face images in the Yale B, CMU PIE and the Weizmann face databases. The eigen faces technique has been used as the baseline algorithm in order to compare the recognition results obtained.

Inter-frame similarity based video transcoding

Multimedia services over wireless networks were made popular by the arrival of smart handhelds. These devices bring heterogeneity to the wireless networks and to the content creation as the content cannot be delivered in its original format due to the difference in the capabilities of these handhelds. An intermediate stage of processing like transcoding is carried out before delivering multimedia content. Under certain scenarios, it is required to convert the content attributes such as bit rate, frame rate, etc. while still retaining the content format to cater to devices with varying capabilities. We propose an algorithm that prioritizes frames taking into account inter frame similarity to perform frame dropping. The priority value based frame dropping in turn aids in delivering better quality video. Our frame priority assignment algorithm is based on uniform distribution of dropped frames to minimize jitter and maximizing the distance between two consecutive dropped frames.

Clustering based Transcoding for Congestion Control in Wireless Networks

Multimedia applications such as video conferencing and media streaming require service guarantees and demand strict timing requirements. Limited network resources along with heterogeneous client devices usually act as a limiting factor on the seamless global delivery of multimedia data. The key issue that needs to be addressed in adaptation and delivery of such multimedia data to diverse devices is to strike a balance between usage of network resources and the quality of delivered content. In this paper we propose a transcoding approach for video streaming to heterogeneous mobile client devices over a wireless network with limited resources and bandwidth. Our approach is based on: (a) Clustering client devices based on their capabilities to determine the format of content suitable for them; (b) Adapting content appropriate to the determined content formats by transcoding and achieving reduction in content size by removing redundant information from content in the process of transcoding; (d) Using subscriber distribution and the devices they own to optimize the number of versions of content; and (e) Using transcoding to further reduce the size of content for each device cluster while maintaining the quality of content.