PROCÉDÉ ET DISPOSITIF DE NORMALISATION DE LOTS EN LIGNE, D'APPRENTISSAGE SUR APPAREIL ET D'APPRENTISSAGE CONTINU APPLICABLES AUX APPAREILS MOBILES OU AUX APPAREILS IOT SE RÉFÉRANT EN OUTRE À UN OU PLUSIEURS LOTS PRÉCÉDENTS DESTINÉS À ÊTRE UTILISÉS À DES FINS MILITAIRES, À UN DRONE OU À UN ROBOT ET PROCÉDÉ ET DISPOSITIF DE TEST LES UTILISANT

摘要

A method of online batch normalization, on-device learning, or continual learning which are applicable to mobile devices, loT devices, and the like is provided. The method includes steps of: (a) computing device instructing convolutional layer to acquire k-th batch, and to generate feature maps for k-th batch by applying convolution operations to input images included in k-th batch respectively; and (b) computing device instructing batch normalization layer to calculate adjusted averages and adjusted variations of the feature maps by referring to the feature maps in case k is 1, and the feature maps and previous feature maps, included in at least part of previous batches among previously generated first to (k-1)-th batches in case k is integer from 2 to m, and to apply batch normalization operations to the feature maps. Further, the method may be performed for military purpose, or other devices such as drones, robots.