抄録
We proposed the object detection method based on saliency map by using reference image containing complex background for service robot. In the previous detection method, images which the user prepares in advance contain mostly simple background. However, in order to make robots do the daily tasks, it is desirable for these robots to be able to detect an object by using snapshots which contain a complex background. In order to decrease effects of features in the background, our proposed method classify local features based on saliency from images. This paper shows the efficacy of proposed method and that our service robot detects some object according to the proposed method.
