How Do You Help a Robot to Find a Place? A Supervised Learning Paradigm to Semantically Infer about Places
In this paper a visual place recognition algorithm suitable for semantic inference is presented. It combines place and object classification attributes suitable for the recognition of congested and cluttered scenes. The place learning task is undertaken by a method capable of abstracting appearance information from the places to be memorized. The detected visual features are treated as a bag of words and quantized by a clustering algorithm to form a visual vocabulary of the explored places. Each query image is represented by a consistency histogram spread over the memorized vocabulary. Simultaneously, an object recognition approach based on Hierarchical Temporal Memory network, updates the robot’s belief of its current position exploiting the features of scattered objects within the scene. The input images which are introduced to the network undergo a saliency computation step and are subsequently thresholded based on an entropy metric for detecting multiple objects. The place and object decisions are fused by voting to infer the semantic attributes of a particular place. The efficiency of the proposed framework has been experimentally evaluated on a real dataset and proved capable of accurately recognizing multiple dissimilar places.
Keywordsplace recognition HTM saliency map semantics robot navigation
Unable to display preview. Download preview PDF.
- 2.Torralba, A., Murphy, K., Freeman, W., Rubin, M.: Context-based vision system for place and object recognition. In: International Conference on Computer Vision, ICCV 2003, pp. 273–280. IEEE (2003)Google Scholar
- 4.Filliat, D.: A visual bag of words method for interactive qualitative localization and mapping. In: International Conference on Robotics and Automation, ICRA 2007, pp. 3921–3926. IEEE (2007)Google Scholar
- 5.Fraundorfer, F., Engels, C., Nistér, D.: Topological mapping, localization and navigation using image collections. In: International Conference on Intelligent Robots and Systems, IROS 2007, pp. 3872–3877. IEEE (2007)Google Scholar
- 8.Hawkins, J., Blakeslee, S.: On intelligence: How a new understanding of the brain will lead to the creation of truly intelligent machines. Henry Holt & Company, New York (2004)Google Scholar
- 11.Kostavelis, I., Nalpantidis, L., Gasteratos, A.: Object recognition using saliency maps and htm learning. In: IEEE International Conference on Imaging Systems and Techniques, IST 2012, pp. 528–532. IEEE (2012)Google Scholar
- 12.Csurka, G., Dance, C., Fan, L., Willamowski, J., Bray, C.: Visual categorization with bags of keypoints. In: Workshop on Statistical Learning in Computer Vision, ECCV 2004, vol. 1, p. 22 (2004)Google Scholar
- 13.Chang, C.C., Lin, C.J.: LIBSVM: A library for support vector machines. ACM Transactions on Intelligent Systems and Technology 2, 27:1–27:27 (2011), Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm
- 16.Kostavelis, I., Gasteratos, A.: Cognitive Navigation Dataset, Group of Robotics and Cgnitive Systems (2012), http://robotics.pme.duth.gr/kostavelis/Dataset.html