Fast and Automatic Detection and Segmentation of Unknown Objects

This research focuses on the fast and automatic detection and segmentation of unknown objects in unknown environments. Many existing object detection and segmentation methods assume prior knowledge about the object or human interference. However, an

Fig 1. Probabilites of super pixels belonging to the foreground based
on color, disparity, and plane information respectively. The white
contour depicts the fixated super pixel.
autonomous system operating in the real world will often be confronted with novel objects. To solve this problem, we propose a segmentation approach named Automatic Detection And Segmentation (ADAS).

Object detection

For the detection of objects, we use symmetry, one of the Gestalt principles for figure-ground segregation to detect salient objects in a scene (see here.

Fig 2. The layout of the Markov Random Field.
The labeling of nodes is based on graph cuts.

Object segmentation

From the initial seed, the object is segmented by iteratively applying graph cuts. We base the segmentation on both 2D and 3D cues: color, depth, and plane information (see Fig. 1). Instead of using a standard grid-based representation of the image, we use super pixels. Besides being a more natural representation, the use of super pixels greatly improves the processing time of the graph cuts, and provides more noise-robust color and depth information. The graphical layout of the Markov Random Field is shown in Fig 2.

The results show that both the object-detection as well as the object-segmentation method are successful and outperform existing methods.


  • Kootstra, G., Bergström, N., & Kragic, D. (2010). Fast and Automatic Detection and Segmentation of Unknown Objects. To be presented at IEEE-RAS International Conference on Humanois Robotics (Humanoids 2010), December 6-8, 2010, Nashville, TN. pdf