Category-independent lower layers of a hierarchical spatial model

Mobile robots need a compact and expressive representation of space to operate effectively in real world environments. We present here:

  • The category-independent lower layers of a hierarchical representation of space that is based on laser range data, which has a potential to scale well.
  • A low-level image descriptor, by which the performance of our representation is demonstrated in the context of a room categorization problem.

A large, freely available, Domestic Rooms Dataset has been collected, which is intended for room categorization experiments based on data obtained with a laser range finder.

Spatial Hierarchy of Parts:

A compositional hierarchical representation of space is learned from observations (Figure 1) and is comprised of several layers of parts (Figure 2). We present here the lower layers that are category-independent and are based on unsupervised learning of statistically significant observations, in terms of frequency of occurrence of various shapes in the environment.

Figure 1: Three example images that were created from laser scans acquired at different positions in one of the living rooms of the Domestic Rooms Dataset.
Figure 2: The Spatial Hierarchy of Parts. (a) Layer 1. (b) Layer 2. (c) Layer 3.

The Histogram of Compositions:

Each laser scan is transformed into an image, from which elements of the hierarchy (parts) are inferred. These are used to form several histograms, where each of them corresponds to some particular region of the image. Their concatenation represents the Histogram of Compositions descriptor (Figure 3), which is used as the input for the Support Vector Machine classifier.

Figure 3: The Histogram of Compositions

Two Scenarios for Room Categorization:

Two scenarios for room categorization have been considered. In the first scenario, called exploratory room categorization, the categorization is performed based on a set of laser scans obtained in each room, while in the second scenario, called single-shot room categorization, the categorization is performed based only on a single scan.

Some results:

Using only the lower layers of the hierarchy, we obtain state-of-the-art categorization results on demanding datasets. Table 1 shows some results obtained on our Domestic Rooms Dataset in the exploratory room categorization scenario. Considering single-shot room categorization scenario, experiments have been performed on Freiburg B79 and B101 datasets (Mozos et al. 2005), and also on the well-established COLD dataset (Ullah et al. 2008). See Figure 5 and Table 2 respectively. For more results and details as well as further discussion, please see the on-line paper “Room Categorization Based on a Hierarchical Representation of Space” published at International Journal of Advanced Robotic Systems.

Table 1: Results of room categorization experiments on the Domestic Rooms Dataset.
Figure 4: Categorization results obtained on Freiburg B79 and B101 datasets. Different colors correspond to different predicted categories: dark blue-room, red-corridor, yellow-doorway, light blue-hall.
Table 2: Confusion matrix displaying the results of the room categorization experiments performed on the COLD dataset. Abbreviations denote: CR-Corridor, 2PO-two-person office, SA-stairs area, TL-toilet.


Publications for the topic of Spatial Hierarchy OfParts