Research And Application Of Visual Query Technology Towards Gesture Interaction

The visual query technology towards gesture interaction provides users with a simpler and more natural way of querying.However,the gesture increases the uncertainty of query events by its natural interactive properties.It has become an important issue that how to obtain accurate and effective query tasks from the gesture interaction.In addition,the query results should be converted into data view to provide interactive-level feedback.The existing dimension reduction can not satisfy the reduction efficiency and topology stability at the same time.In view of the above two key issues,the main work of this thesis is as follows:(1)This thesis summarizes and analyzes the new problems brought by the introduction of gesture interaction to the traditional visual query,and higher requirements of real-time response.The importance of the decision strategy of query events and the technology of dimension reduction for query results are expounded.In addition,this thesis introduces the recent research status of the key theories and the application of query technology for gesture interaction.(2)In order to provide nearly real-time feedback during the process of gesture interaction,a query strategy based on features of multi-touch interactions is proposed.A query process can be dynamically divided based on spatio-temporal continuity and velocity characteristics in the process of touching interaction.To build new queries more efficiently,the interaction context is redefined as a collection of touch interaction features and operating objects.And an optimization method based on the comparison of context parameters is proposed to avoid the continuous and repeated queries.By collecting the real touching behavior data,the experimental results show that the proposed method has a better comprehensive performance,compared with SnaptoQuery?touchViz and Binned aggregation.(3)In order to quickly obtain the low-dimensional description information of query results in visual space,an isometric mapping algorithm with global adaptability is proposed,which is called GA-Isomap.Based on the basic framework of traditional isometric mapping and minimum connected graphs,a method of increamental constructing neighborhood graphs and selecting regional landmarks are proposed.And a mapping calculation method based on bilevel graph is proposed.Landmark points are not only used to reduce the computational workload,but also used to improve the data topology stability by relative position relations.Compared with original Isomap,L-Isomap,Isomap with dynamic neighbor and Isomap with NC,simulation results show that the GA-Isomap algorithm can effectively balance the data topology stability and operational efficiency in data visualization.(4)Based on the above research results and the current popular interactive visualization framework D3,using air quality query as a practical application background,a prototype system based on gesture interaction is designed and implemented.The air quality data can be queried by date,pollutant category,index and location,and visualized in the form of scatter bubble,line and histogram.