Beyond histograms: why learned structure-preserving descriptors outperform HOG.
Thomas Guthier, Volker Willert, Julian P. Eggert · 2014
Statistical image descriptors based on histograms (e.g. SIFT [1], HOG [2]) are widely used in image processing, because they are fast and simple methods with high classification performance. However, they discard the local spatial topology and thus lose discriminative information contained in the image. We discuss the relations between HOG and VNMF descriptors, i.e. structure free histograms ver-sus learned structure-preserving patterns. VNMF is a shift-invariant, sparse, non-negative unsupervised learning algorithm [8, 9, 5], that provides a distinct decom-position of the input into its parts. The VNMF descriptor outperforms the statistical HOG descriptor, because it preserves spatial topology leading to better classifica-tion results on real-world human action recognition benchmarks [11, 12]. 1