%0 Conference Proceedings
%T Exploration Strategies for Incremental Learning of Object-Based Visual Saliency
%+ Robotique et Vision (RV)
%+ Flowing Epigenetic Robots and Systems (Flowers)
%+ Thales Research and Technology [Palaiseau]
%A Craye, Céline
%A Filliat, David
%A Goudou, Jean-François
%< avec comité de lecture
%B Joint IEEE International Conference Developmental Learning and Epigenetic Robotics (ICDL-EPIROB)
%C Providence, United States
%8 2015-08-13
%D 2015
%Z Computer Science [cs]/Robotics [cs.RO]Conference papers
%X Searching for objects in an indoor environment can be drastically improved if a task-specific visual saliency is available. We describe a method to learn such an object-based visual saliency in an intrinsically motivated way using an environment exploration mechanism. We first define saliency in a geometrical manner and use this definition to discover salient elements given an attentive but costly observation of the environment. These elements are used to train a fast classifier that predicts salient objects given large-scale visual features. In order to get a better and faster learning, we use intrinsic motivation to drive our observation selection, based on uncertainty and novelty detection. Our approach has been tested on RGB-D images, is real-time, and outperforms several state-of-the-art methods in the case of indoor object detection.
%G English
%2 https://hal.science/hal-01170532/document
%2 https://hal.science/hal-01170532/file/ICDL_celine_craye_final.pdf
%L hal-01170532
%U https://hal.science/hal-01170532
%~ ENSTA
%~ INRIA
%~ INRIA-BORDEAUX
%~ INRIA_TEST
%~ TESTALAIN1
%~ ENSTA_U2IS
%~ TESTBORDEAUX
%~ TESTBORDEAUX2
%~ INRIA2
%~ UNIV-PARIS-SACLAY
%~ ENSTA-SACLAY