Bhavan Gandhi

Intelligent multimedia content management on mobile devices

The increasing capability of mobile handsets, advancing multimedia processing, improving codec technology (e.g., mp3, JPEG, MPEG-4 SP, H.264), and bigger communication pipes have the potential for providing increasing volumes of multimedia data to mobile users. The problem of intelligently managing multimedia content is becoming increasingly prevalent in this space. For example, applications that archive and retrieve personal content as well as those that search and stream commercial content over wired and wireless channels need to provide a compelling user experience while transparently and efficiently handling the vast amounts of underlying data. This work provides a user-intuitive, standards-based approach for managing multimedia content on mobile handsets.

Optimal video summarization with a bit budget constraint

The need for video summarization originates primarily from a viewing time or a bit budget constraint. A shorter version of the original video sequence is desirable in a number of applications. Clearly, a shorter version is also necessary in applications where storage, communication bandwidth and/or power are limited, which translates into a bit budget constraint. Our work is based on a bit rate-summary distortion optimization formulation. New metrics for video summary distortion are introduced. An optimal algorithm based on Lagrangian relaxation and dynamic programming is presented.

Perceptual Feature Selection for Semantic Image Classification

Content-based image retrieval has become an indispensable tool for managing the rapidly growing collections of digital images. The goal is to organize the contents semantically, according to meaningful categories. In recent papers we introduced a new approach for semantic image classification that relies on the adaptive perceptual color-texture segmentation algorithm proposed by Chen et al. This algorithm combines knowledge of human perception and signal characteristics to segment natural scenes into perceptually uniform regions. The resulting segments can be classified into semantic categories using region-wide features as medium level descriptors. Such descriptors are the key to bridging the gap between low-level image primitives and high-level image semantics. The segment classification is based on linear discriminant analysis techniques. In this paper, we examine the classification performance (precision and recall rates) when different sets of region-wide features are used. These include different color composition features, spatial texture, and segment location. We demonstrate the effectiveness of the proposed techniques on a database that includes 9000 segments from approximately 2500 photographs of natural scenes.

Fast video shot retrieval by trace geometry matching in principal component space

Content-based video retrieval technology holds the key to the efficient management and sharing of video content from different sources, in different scales, across different platforms, and shared over different communication channels. In this work we present a fast retrieval algorithm based on matching the geometry of video sequence traces in the principal component space. Techniques to address scale (spatial and temporal) issues, as well as, noise and other possible distortions, such as frame dropping, are discussed. Experimental results demonstrate the effectiveness of the proposed approach.

Rate-distortion optimal video summarization: a dynamic programming solution

The need for video summarization originates primarily from a viewing time constraint. A shorter version of the original video sequence is desirable in a number of applications. Clearly, a shorter version is also necessary in applications where storage, communication bandwidth and/or power are limited. Our work is based on a temporal rate-distortion optimization formulation for optimal summary generation. New metrics for video summary distortion are introduced. Optimal algorithms based on dynamic programming are presented along with the results from heuristic algorithms that can produce near optimal results in real time.

A Compressed XML Schema Representation for Metadata Processing in Mobile Environments

Efficient packaging and communication of metadata are critical in multimedia communications to achieve seamless mobility. The XML schema compression proposed here comprises a method for decomposing an XML schema into a sequence of atomic elements. This representation reorganizes the given XML schema with the threefold purpose of facilitating dynamic schema switching and reconfiguration of metadata decoders, increasing the efficiency of binary metadata decoding, and storing XML schemas in binary format. As a result, schemas can be efficiently transmitted over congested networks and stored in devices with limited resources. Since each schema may represent the metadata structure for a different type of media format, this allows a system to support different profiles (schemas) and media formats, enabling dynamically reconfigurable metadata encoders and decoders

A Method to Compress Schema-Based XML Metadata for Mobile Environments

Efficient and error-free packaging and communication of XML metadata are critical in achieving seamless mobility. This paper describes an XML binary compression technology using both schema knowledge (to separate the structural and value portions of XML metadata) and suitable compression techniques for different metadata portions. The compression results are superior to comparable XML compression mechanisms. In addition, the compression efficiency has much less dependency on the metadata size. This compression technology also goes beyond existing metadata compression technologies by providing a direct mapping of XML metadata into a database interface so that the decoded metadata can be directly queried for retrieval. Finally, error resilient features are an integral part of the compressed metadata bitstream so that data corrupted during transmissions can be detected within isolated groups and exclusively retransmitted for efficient network resource utilization.