PROJECT TITLE :
Predicting Detection Performance on Security X-Ray Images as a Function of Image Quality
Research into how image quality impacts work performance is a hot topic in many industries. The security X-ray imaging literature lacks investigations of this nature that have been conducted in the medical imaging sector. Bomb technicians use portable X-ray devices to detect improvised explosive device components, and this paper examines how picture quality affects their ability to identify these components. Objective techniques for predicting bomb technician detection performance have been built using a newly established NIST-LIVE Task Performance Database. Image quality indicators (IQIs), as well as perceptually relevant natural scene statistics (NSS)-based metrics, have been extensively used in the prediction algorithms for the quality of visible light images. Expert bomb technicians' ability to identify dangers is predicted by these metrics, which we demonstrate can quantify the perceptual severity of degradations. Even greater task performance prediction can be achieved by combining NSS and IQI-based measurements. This is the first NSS-based model for security X-ray pictures, which we call the Quality Inspectors of X-ray Images (QUIX). We also built a new set of statistical job prediction models. We also demonstrate that standard IQI metric values on distorted X-ray pictures may be successfully predicted using QUIX.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here