Closed-loop object recognition using reinforcement learning

Current computer vision systems whose basic methodology is open-loop or filter type typically use image segmentation followed by object recognition algorithms. These systems are not robust for most real-world applications. In contrast, the system presented here achieves robust performance by using reinforcement learning to induce a mapping from input images to corresponding segmentation parameters. This is accomplished by using the confidence level of model matching as a reinforcement signal for a team of learning automata to search for segmentation parameters during training. The use of the recognition algorithm as part of the evaluation function for image segmentation gives rise to significant improvement of the system performance by automatic generation of recognition strategies. The system is verified through experiments on sequences of color images with varying external conditions.

[1]  Keith Price,et al.  Picture Segmentation Using a Recursive Region Splitting Method , 1998 .

[2]  Takeo Kanade,et al.  Recursive region segmentation by analysis of histograms , 1982, ICASSP.

[3]  Steven A. Shafer,et al.  The Phoenix Image Segmentation System: Description and Evaluation , 1982 .

[4]  Takayuki Ito,et al.  Neocognitron: A neural network model for a mechanism of visual pattern recognition , 1983, IEEE Transactions on Systems, Man, and Cybernetics.

[5]  Charles R. Dyer,et al.  Model-based recognition in robot vision , 1986, CSUR.

[6]  Bir Bhanu,et al.  Recognition of occluded objects: A cluster-structure algorithm , 1987, Pattern Recognit..

[7]  Kumpati S. Narendra,et al.  Learning automata - an introduction , 1989 .

[8]  A. Barto,et al.  Learning and Sequential Decision Making , 1989 .

[9]  Lawrence D. Jackel,et al.  Backpropagation Applied to Handwritten Zip Code Recognition , 1989, Neural Computation.

[10]  Jing Peng,et al.  Function Optimization using Connectionist Reinforcement Learning Algorithms , 1991 .

[11]  David Chapman,et al.  Intermediate Vision: Architecture, Implementation, and Use , 1992, Cogn. Sci..

[12]  Pascal Fua,et al.  Computational strategies for object recognition , 1992, CSUR.

[13]  David Chapman,et al.  Intermediate vision: Architecture, implementation, and use☆ , 1992 .

[14]  F. Girosi,et al.  Some Extensions of the K-Means Algorithm for Image Segmentation and Pattern Classification , 1993 .

[15]  B. Bhanu,et al.  Image understanding research for automatic target recognition , 1993, IEEE Aerospace and Electronic Systems Magazine.

[16]  Ron Kohavi,et al.  Irrelevant Features and the Subset Selection Problem , 1994, ICML.

[17]  B. Bhanu,et al.  Adaptive image segmentation using genetic and hybrid search methods , 1995, IEEE Transactions on Aerospace and Electronic Systems.

[18]  Bir Bhanu,et al.  Adaptive image segmentation using a genetic algorithm , 1989, IEEE Transactions on Systems, Man, and Cybernetics.

[19]  Bir Bhanu,et al.  Delayed reinforcement learning for closed-loop object recognition , 1996, Proceedings of 13th International Conference on Pattern Recognition.

[20]  Visvanathan Ramesh,et al.  Performance characterization of image understanding algorithms , 1996 .

[21]  Kannan,et al.  ON IMAGE SEGMENTATION TECHNIQUES , 2022 .