Visual analysis of large dental imaging data in caries research

Abstract

With dental imaging data acquired at unprecedented speed and resolution, traditional serial image processing and single-node storage need to be re-examined in a “BigData” context. Furthermore, most previous dental computing has focused on the actual imaging acquisition and image analysis tools, while much less research has focused on enabling caries assessment via visual analysis of large dental imaging data. In this paper we present DENVIS, an end-to-end solution for cariologists to manage, mine, visualize, and analyze large dental imaging data for investigative carious lesion studies. DENVIS consists of two main parts: data driven image analysis modules triggered by imaging data acquisition that exploit parallel MapReduce tasks and ingest visualization archive into a distributed NoSQL store, and user driven modules that allow investigative analysis at run time. DENVIS has seen early use by our collaborators in oral health research, where our system has been used to pose and answer domain-specific questions for quantitative assessment of dynamic carious lesion activities.