Ensemble based constrained-optimization extreme learning machine for landmark recognition

Landmark recognition attracts great concerns in recent years due to its extensive applications in mobile terminals. An effective recognition system with high recognition accuracy and fast response speed is highly desired by users. In this paper, we propose an ensemble based constrained-optimization extreme learning machine (CO-ELM) combining with the spatial pyramid kernel based bag-of-words (SPK-BoW) method for landmark recognition. The recent SPK-BoW method is employed for feature extraction and representation due to its effectiveness in exploiting the spatial layout information for landmark images. To enhance the recognition performance and accelerate the data training and testing speed, the voting based CO-ELM (VCO-ELM) with multiple network ensembles is proposed as the classifier. Experiments on two real-world landmark datasets show that the proposed VCO-ELM algorithm outperforms the original CO-ELM and support vector machine (SVM) in general.