Finding objects at indoor environment combined with depth information

Abstract

To detect and localize objects in a scene is an essential step for many computer vision tasks. Many efforts have been done for detecting and localizing category-specific ojects. However, only few works focused on the generic objectness measure which is more common and important than category-specific object detection. Base on an existing method, in this paper, a novel method by combining a new cue, the depth information, is proposed for detecting and localizing generic objects in the indoor scenes. Through our experiments, we found that by adding depth information cue the performance of detecting and localizing will be better (especially for the closed objects) than the original method. Finally, a method for selecting bounding box which contains a possible object is also introduced and the result is promising by being shown in our experiments.