Validating the quality of millimeter-wave images input to deep-learning-based threat detection systems
Jack L. Glover, Praful Gupta, Marius B. Facktor, Alan C. Bovik
Deep-learning-based automatic image decision systems are increasingly being relied on to interpret imagery that was previously only viewed and analyzed by humans. Here, we present a real-time method of validating the quality of images input to image decision systems. We focus on the detection of concealed contraband in millimeter-wave (MMW) images of screened people, but the method is general enough to be useful for other applications, such as medical image analysis systems. In applications of such critical importance, it is imperative that automatic target recognition (ATR) algorithms behave predictably and robustly. For example, a MMW system deployed in an airport could suffer from changes in image quality due to a variety of factors, for example, partial hardware malfunctions, excessive vibration, or lack of maintenance or calibration. In such scenarios, it would be desirable to be able to detect changes in image quality in real-time. We investigate the performance of a deep-learning-based ATR when fed with variable quality input images. We describe a first-of-its-kind method to validate the quality of images input to an ATR. The real-time method uses statistical measurements intrinsic to natural images to assess the similarity between an input image and a set of training images. We show, through multiple experiments, that recognition performance is significantly worse on images with quality that is significantly different than the training set. Interestingly, this also appears to hold when the ATR processes images of better quality than the training set. The method is successfully demonstrated as a training-free validation tool for ATR algorithms using two state-of-the-art deep-learning architectures.
Proceedings of the SPIE 11729, Automatic Target Recognition XXXI
April 12-17, 2021
Online only due to COVID, MD, US
SPIE Defense + Commercial Sensing
security imaging, object detection, image quality, deep learning, natural scene statistics, millimeter wave SAR
, Gupta, P.
, Facktor, M.
and Bovik, A.
Validating the quality of millimeter-wave images input to deep-learning-based threat detection systems, Proceedings of the SPIE 11729, Automatic Target Recognition XXXI, Online only due to COVID, MD, US
(Accessed September 30, 2023)