Interactive fast random access, retrieval, and navigation of large datasets
Zihong Fan · 2015
This research is motivated by two important trends. First, more than ever before large amounts of data and information are being accessed through mobile devices such as smart phones, tablet computers, book readers, etc. Second, significant amounts of complex and high-volume information (e.g. maps, medical images, scientific datasets, virtual museums, etc.) are now available over networks. In addition to entertainment applications (e.g. video sharing), a significant driver of traffic over networks is likely to come from professional applications. Such applications, like those that might provide doctors with pervasive access to medical information, will demand high quality performance according to a variety of different metrics, such as latency, resolution, interactivity, and perceptual quality. ❧ We have proposed a novel system for interactive, fast, and random access and navigation of large datasets. Of particular interest in our system is the random retrieval of lower dimensional data from high dimensional datasets. The system makes it possible to allow limited memory mobile devices to quickly access complex large datasets over low-bandwidth connections. This approach can reduce the transmission rate dramatically (by factor of 2) and also improve the level of interactive navigation. Both 2D and 3D systems, referring to the dimensionality of the datasets that may be accessed, for fast and random access have been developed and each system is composed of a tiling scheme, linear searching algorithm, compression methodology, mapping algorithm, and reconstruction scheme. The mapping algorithms can also be applied to various other applications and areas. We also have proposed models for parameter selection and system optimization, tools to both analyze and quantify the benefits of our proposed re-mapping algorithm as well as wavelet based approach 3D system. This work offers both theoretical and practical contributions and can be used in a wide-array of exciting future applications.