Active object recognition by view integration and reinforcement learning

被引:79
作者
Paletta, L [1 ]
Pinz, A [1 ]
机构
[1] Graz Univ Technol, Inst Comp Graph & Vis, A-8010 Graz, Austria
基金
奥地利科学基金会;
关键词
active recognition; reinforcement learning; information fusion; viewpoint planning;
D O I
10.1016/S0921-8890(99)00079-2
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
A mobile agent with the task to classify its sensor pattern has to cope with ambiguous information. Active recognition of three-dimensional objects involves the observer in a search for discriminative evidence, e.g., by change of its viewpoint. This paper defines the recognition process as a sequential decision problem with the objective to disambiguate initial object hypotheses. Reinforcement learning provides then an efficient method to autonomously develop near-optimal decision strategies in terms of sensorimotor mappings. The proposed system learns object models from visual appearance and uses a radial basis function (RBF) network for a probabilistic interpretation of the two-dimensional views. The information gain in fusing successive object hypotheses provides a utility measure to reinforce actions leading to discriminative viewpoints. The system is verified in experiments with 16 objects and two degrees of freedom in sensor motion. Crucial improvements in performance are gained using the learned in contrast to random camera placements. (C) 2000 Elsevier Science B.V. All rights reserved.
引用
收藏
页码:71 / 86
页数:16
相关论文
共 36 条
  • [1] ALOIMONOS Y, 1990, P 10 INT C PATT REC, P346
  • [2] [Anonymous], 1990, TASK DIRECTED SENSOR
  • [3] Baird L, 1995, MACHINE LEARNING P 1, P30
  • [4] PRINCIPLES OF ANIMATE VISION
    BALLARD, DH
    BROWN, CM
    [J]. CVGIP-IMAGE UNDERSTANDING, 1992, 56 (01): : 3 - 21
  • [5] BANDERA C, 1996, P 13 INT C MACH LEAR, P20
  • [6] BOROTSCHNIG H, 1998, P BRIT MACH VIS C, P629
  • [7] Fast-learning VIEWNET architectures for recognizing three-dimensional objects from multiple two-dimensional views
    Bradski, G
    Grossberg, S
    [J]. NEURAL NETWORKS, 1995, 8 (7-8) : 1053 - 1080
  • [8] Bridle J. S., 1990, Neurocomputing, Algorithms, Architectures and Applications. Proceedings of the NATO Advanced Research Workshop, P227
  • [9] PSYCHOPHYSICAL SUPPORT FOR A 2-DIMENSIONAL VIEW INTERPOLATION THEORY OF OBJECT RECOGNITION
    BULTHOFF, HH
    EDELMAN, S
    [J]. PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 1992, 89 (01) : 60 - 64
  • [10] Autonomous recognition: Driven by ambiguity
    Callari, FG
    Ferrie, FP
    [J]. 1996 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, PROCEEDINGS, 1996, : 701 - 707