Selecting Local Region Descriptors with a Genetic Algorithm for Real-World Place Recognition
The basic problem for a mobile vision system is determining where it is located within the world. In this paper, a recognition system is presented that is capable of identifying known places such as rooms and corridors. The system relies on a bag of features approach using locally prominent image regions. Real-world locations are modeled using a mixture of Gaussians representation, thus allowing for a multimodal scene characterization. Local regions are represented by a set of 108 statistical descriptors computed from different modes of information. From this set the system needs to determine which subset of descriptors captures regularities between image regions of the same location, and also discriminates between regions of different places. A genetic algorithm is used to solve this selection task, using a fitness measure that promotes: 1) a high classification accuracy; 2) the selection of a minimal subset of descriptors; and 3) a high separation among place models. The approach is tested on two real world examples: a) using a sequence of still images with 4 different locations; and b) a sequence that contains 8 different locations. Results confirm the ability of the system to identify previously seen places in a real-world setting.
Unable to display preview. Download preview PDF.
- 1.Faugeras, O.: Three-Dimensional Computer Vision (Artificial Intelligence). The MIT Press, Cambridge (1993)Google Scholar
- 2.Sivic, J., Russell, B.C., Efros, A.A., Zisserman, A., Freeman, W.T.: Discovering objects and their localization in images. In: Proceedings of the 10th IEEE International Conference on Computer Vision (ICCV 2005), Beijing, China, 17-20 October 2005, vol. 1, pp. 370–377. IEEE Computer Society, Los Alamitos (2005)CrossRefGoogle Scholar
- 3.Willamowski, J., Arregui, D., Csurka, G., Dance, C., Fan, L.: Categorizing nine visual classes using local appearance descriptors. In: Proceedings of ICPR 2004, Workshop on Learning for Adaptable Visual Systems, Cambridge, United Kingdom, 23-26 August 2004, IEEE Computer Society, Los Alamitos (2004)Google Scholar
- 4.Trujillo, L., Olague, G., de Vega, F.F., Lutton, E.: Evolutionary feature selection for probabilistic object recognition, novel object detection and object saliency estimation using gmms. In: BMVC 2003: Proceedings of the 18th British Machine Vision Conference. British Machine Vision Association, vol. 2, pp. 630–639 (2007)Google Scholar
- 6.Trujillo, L., Olague, G.: Using evolution to learn how to perform interest point detection. In: Proceedings of ICPR 2006, Hong Kong, China, 20-24 August 2006, vol. 1, pp. 211–214. IEEE Computer Society, Los Alamitos (2006)Google Scholar
- 8.Torralba, A., Murphy, K.P., Freeman, W.T., Rubin, M.A.: Context-based vision system for place and object recognition. In: ICCV 2003: Proceedings of the Ninth IEEE International Conference on Computer Vision, Washington, DC, USA, p. 273. IEEE Computer Society, Los Alamitos (2003)CrossRefGoogle Scholar
- 14.Fisher, R.A.: The use of multiple measurements in taxonomic problems. Annals of Eugenics 7, 179–188 (1936)Google Scholar