A way of the robust acceleration optimization of image identification of X-ray machine used in airport security check based on the bag of words database model

In airport security check, the demands of the accuracy of image identification of X-ray machine operators have become higher and higher. The different positions of items in the conveyor make the images shown in the computer displays different, which brings difficulty to the accurate identification. This article makes an analysis of the images and puts forward a way of robust acceleration optimization against the classical bag of words model (which has some flaws and needs to be improved). This new way can describe precisely the graphical features of the visual dictionary produced by visual images of Xray machines, resist the influence of the complicated location and background information and categorize the information that is put in the sorters of support vector machines (SVM). Through experiments and analysis, it is proved that this way can increase the accuracy of the operators’ graphical identification and achieve a good effect with few experimental images, which means it can increase the accuracy and the efficiency of the operators’ identification of difficult images.