Gesture Recognition Based on Sparse Reconstruction

Due to the variety of feature extractions and classifiers, many different algorithms have been proposed for gesture recognition. In this paper, we work to increase the recognition performance in terms of recognition rate and execution time by using recently proposed modified sparse representation classifier based on intensity of images. Sparsity based classifier is compared with two conventional ones as K-nearest neighbor and random forest classifiers on gesture recognition. Simulation results showed that, our recognition algorithm based on sparsity has a higher performance than that of the others for both recognition rate and execution time.