RL-IAC: An exploration policy for online saliency learning on an autonomous mobile robot

Abstract

In the context of visual object search and localization, saliency maps provide an efficient way to find object candidates in images. Unlike most approaches, we propose a way to learn saliency maps directly on a robot, by exploring the environment, discovering salient objects using geometric cues, and learning their visual aspects. More importantly, we provide an autonomous exploration strategy able to drive the robot for the task of learning saliency. For that, we describe the Reinforcement Learning-Intelligent Adaptive Curiosity algorithm (RL-IAC), a mechanism based on IAC (Intelligent Adaptive Curiosity) able to guide the robot through areas of the space where learning progress is high, while minimizing the time spent to move in its environment without learning. We demonstrate first that our saliency approach is an efficient tool to generate relevant object boxes proposal in the input image and significantly outperforms the state-of-the-art EdgeBoxes algorithm. Second, we show that RL-IAC can drastically decrease the required time for learning saliency compared to random exploration.

Exploring to learn visual saliency: The RL-IAC approach

“…This article extends Craye et al [19] by providing more technical details and experimental results. It also presents a new type of feature extractor based on convolutional neural networks, and a method to automatically and incrementally obtain a navigation graph required by the RL-IAC algorithm.The CNN feature extraction approach was already presented in a previous work [18], but the evaluation was carried out on a different dataset, and the use of RL-IAC was not investigated.…”

“…In this article, we consider a mobile robot exploring its environment while building a model of visual saliency enhancing objects of interest. Based upon our previous work [19], we present a system ( Fig. 1) able to:…”

BioVision: A Biomimetics Platform for Intrinsically Motivated Visual Saliency Learning

“…We instead use an object detector based on the depth-map to this end. This depth-based object detection has been described in previous publications that one could refer to for more details [14], [15]. Apart from this, the saliency learning process is exactly the same.…”

“…We therefore apply the superpixels approach presented in Section IV-D to increase the resolution of these maps. To evaluate our feature extractor versus the one proposed in previous work [15], we generate saliency maps from both the CNN-based feature extractor (denoted as ISL here), and the former feature extractor (denoted as ISLMake3D) that was used in [15].…”

“…However, the IAC algorithm was not applied yet for this kind of setups, and the evaluation was conducted on publicly available datasets rather than on the BioVision platform. In Craye et al [15], we investigated the use of IAC on a mobile robot, where the goal was to displace the robots in different rooms of a building to efficiently learn visual saliency. We aim to demonstrate in this paper that the IAC mechanism is also applicable to a foveated system by taking advantage of the foveal and contextual views.…”