अमूर्त

A study of X-ray machine image local semantic features extraction model based on Bag-of-Words for airport security

Ning Zhang, Jinfu Zhu

The aviation security at the airport has been faced with increasingly severe situations since the 9-11 event. It’s of utmost importance to train airport X-ray machine screeners image recognition competency. So they can prevent terrorists from bringing dangerous articles in their carry-on or checked bags. However, usually the luggages are placed in different positions and the density & volume of articles differ greatly. As a result, dangerous articles show a variety of X-ray image features. It’s easy for the confused screeners to miss or incorrectly detect dangerous articles. This has been a hidden danger for civil aviation safety. For image recognition improvement, the researcher analyzed the visual semantics of dangerous goods images and applied a local semantic features extraction method. After classification and summarization, the method was used to train the screeners for particular image recognition. The comparison showed the improved accuracy and efficiency of image recognition for the screeners and demonstrated a satisfactory effect