Accurate measurement of smoking behaviors and exposures may become critical as FDA implements its authority to regulate tobacco products, and in particular proposes product standards and considers approval of modified risk products. Recently, digital image analysis systems have been developed that can identify the blocking of filter vents on spent cigarette filters with high accuracy and can also estimate the degree of smoker compensation. There is strong potential for such digital imaging systems to unobtrusively infer a rich set of smoker topography variables as well as a smoker's exposure to toxins. The existing literature base on filter-based assays is growing and points to growing prominence and applicability of these approaches to important research questions. We propose to modify existing analysis systems and conduct a series of studies to test their reliability, validity, and applicability for determining mouth- level cigarette smoke exposure by examining tar stains on cigarette filter butts.
The Specific Aims are designed to provide rigorous evaluations of the utility of the systems, from establishing prediction equations using machine-smoked cigarettes to cross- validation in human-smoked samples against other established measures of mouth-level exposure. At the conclusion of this project, we plan to release a validated suite of software to the research community to support external verification of digital image analysis for smoking-related research.
Accurately and simply measuring smoking behavior and smoke exposure is important, and digital image analysis systems have been developed that can estimate smoke intake and identify the blocking of filter vents on spent cigarette filters. The proposed project will modify existing analysis systems and conduct a series of studies to test their reliability, validity, and applicability for determining mouth-level cigarette smoke exposure. We plan to make a suite of software publicly available to allow others to apply this technology in their own research.
|Mendrik, Adriënne M; Vincken, Koen L; Kuijf, Hugo J et al. (2015) MRBrainS Challenge: Online Evaluation Framework for Brain Image Segmentation in 3T MRI Scans. Comput Intell Neurosci 2015:813696|