An Ensemble Learning Method Based on Deep Neural and Pca-Based Svm Network for Baggage Threat and Smoke Recognition

Ensemble learning methods are emerging as one of the most widely used image classification and object recognition techniques. They have the advantage of achieving higher classification and recognition performance compared to single models while being easy to implement in various applications. This work proposes an ensemble-based classification network that leverages the recognition performance of both threat and smoke detection tasks. The method utilizes ensemble learning of a deep Convolutional Neural Network (CNN) combined with a Principal Component Analysis (PCA)-based Support Vector Machine (SVM) classifier. Comparisons to several single model classifiers and other state-of-the-art methods and the proposed methods were carried out. The proposed method showed superior performance in comparison for the problems of baggage X-ray imagery classification and smoke recognition.