Statistical Edge Detection of Concealed Weapons Using Artificial Neural Networks

摘要: A novel edge detector has been developed that utilises statistical masks and neural networks for the optimal detection of edges over a wide range image types. The failure many common techniques observed when analysing concealed weapons X-ray images, biomedical images or with significant levels noise, clutter texture. This technique is based on filter uses two-sample tests to evaluate any local texture differences by applying pixel region mask (or kernel) analyse properties region. type greatly expanded from previous work Bowring et al. 1 process further enhanced combined multiple scale tests, Artificial Neural Networks (ANN) trained classify different Through use we can combine output results several scales into one detector. Furthermore allow combination two sample varying (for example; mean based, variance distribution based). both allows response variety masks. From this produce optimum are presented.