Rapid advances in optical instrumentation, high-speed cameras, and fluorescent probes have spurred tremendous growth in the volume of biomolecular imaging data. Various optical imaging modalities are used for probing biological systems in vivo and in vitro. These include traditional two-dimensional imaging, three-dimensional confocal imaging, time-lapse imaging, and multispectral imaging. Many applications require a combination of these imaging modalities, which gives rise to huge data sets. However, lack of powerful information extraction and quantitative analysis tools poses a major hindrance to exploiting the full potential of the information content of these data. In particular, automated extraction of semantic information from multimodality imaging data, crucial for understanding biological processes, poses unique challenges. Information extraction from large sets of biomolecular imaging data requires modeling at multiple levels of detail to allow not only quantitative analysis but also interpretation and extraction of high-level semantic information. In this paper, we survey the state of the art in the area of information extraction and automated analysis tools for in vivo and in vitro biomolecular imaging. The modeling and knowledge extraction for these data require sophisticated image processing and machine learning techniques, as well as formalisms for information extraction and knowledge management. Development of such tools has the potential to significantly improve biological discovery and drug development processes.

Conventional Association Rule Mining algorithms require a user-specified minimum support threshold to mine the frequent itemsets (FIs). However, selecting such an optimum support threshold, to find the desired resulting FIs, is a tough proposition. There will be too many or too few resulting FIs if the threshold is not selected appropriately. Therefore, it is more desirable to have a solution that allows the users to control the amount of results. In this paper, we propose an efficient implementation for a constraint based FI mining problem, i.e. to mine N k-itemsets with the highest support for k up to a certain kmax value. The results thus obtained are called N-most interesting itemsets [1]. We have devised and implemented two new efficient algorithms, named NFOLD-growth and LOOPBACKNFOLD-growth, for the above-mentioned problem. Our implementation is based on a novel data structure, SOTrieIT (Support-Ordered Trie Itemset), proposed recently in literature [2]. Our simulations show that our algorithms significantly outperform the previous solutions for the same problem in terms of execution time. In addition to being support independent and incremental, the algorithms are also more space-efficient than the previous algorithms.

We propose and evaluate techniques to manage disk storage systems for multimedia document servers with constrained I/O resources. We also evaluate the requirements for the main memory to maintain desired levels of quality of presentation (QoP) for multiple users. This is achieved by attempting to minimize the number of tardy frames in end-to-end delivery of multimedia data. Inter-media and intra-media temporal synchronization at the user-interaction level are essential concerns for multimedia servers. Scheduling of multimedia frames has direct impact on both intra-stream and inter-stream multimedia synchronization. In order to effect such synchronization, we propose several O(n log n+mn) heuristics to schedule multimedia frames between the disks and destination memories of the multimedia systems, where m is the number of logical I/O channels and n is the number of frames of multimedia data. We study the performance of these heuristics via simulations and show the tradeoffs between the system resources and QoP parameters.