[Abstract] In computer vision, current feature extraction techniques generate high dimensional data. Both convolutional neural networks and traditional approaches like keypoint detectors are used as extractors of high-level ...