Abstract

With RadSpeech, we aim to build the next generation of intelligent, scalable, and user-friendly semantic search interfaces for the medical imaging domain, based on semantic technologies. Ontology-based knowledge representation is used not only for the image contents, but also for the complex natural language understanding and dialogue management process. This demo shows a speechbased annotation system for radiology images and focuses on a new and effective way to annotate medical image regions with a specific medical, structured, diagnosis while using speech and pointing gestures on the go.

In this paper, we present DynaQ, our prototypical inquiry system to explore the personal information space that supports the user with the help of the searching paradigm Orienteering, achieved by using dynamic queries.Orienteering is typically characterized by relatively small steps following one after another. In contrast to searching through keywords, Orienteering offers several advantages: reduction of cognitive load, sensitivity of the environment, and a better understanding of the results.Dynamic Queries enables the user to specify search queries in a dynamic way to get immediate feedback on changes inside data sets that fulfill the query criteria I. INTRODUCTIONIt is not only the scale of the internet that grows rapidly. Also the information on personal computers is reaching dimensions hard to manage. Over years, numerous documents accumulate at the hard disks, e.g. tutorials, proposals, contracts, documentations, announcements, and thousands of emails. The search for the appropriate information/document with common methods results in huge costs for companies as shown by the investigation of the IWD in 2003 [1]:Costs of the search for documents: Every employee spends on the average 2,44 hours per week for searching documents. This costs 3,74 Mio c/year for a company with 1.000 employees. How does such a search looks like? Nearly every tool follows the same strategy: the keyword search with subsequent presentation of the documents found on a list. Depending on the tool used further constrictions on the searching space can be performed, e.g. documents that need less space than 1 MB. Unfortunately, this does not comply with the natural human searching attitude.Current research of the MIT shows how humans perform personal motivated searches inside their emails and files/documents [2]. For this purpose, over 150 graduated MIT students were interviewed. The researchers detected that the attidutes for searching rarely cover the classical keyword search. In fact, the participants navigated towards their goals with tiny local steps, by the help of their context knowledge. This was also performed in case they had known exactly what they searched for. This searching strategy was named "Orienteering".Thus, Orienteering specifies a searching attitude where humans satisfy their current need for information by a sequence

No abstract