Fabric Image Edge Detection Based on Octonion and Echo State Networks

Abstract:

Based on synthesizing octonions and the R, G, B components of color image , a color image edge detection algorithm is proposed ,combining octonions with Echo State Networks(ESN) . The judged pixel and its eight-neighbourhood mean is used to calculate vector product as the feature vectors of image edge. Then the ESN is trained with training samples and the trained ESN is directly used for edge detection. Experiments show that this method has good effect of fabric edge detection with a stronger ability to keep the details.

• Periodical: Applied Mechanics and Materials (Volumes 519-520)
• Pages: 714-718
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.519-520.714
• Online since: February 2014
• Authors: Jian Cheng Chen*, Ang Yan Tu
• Keywords: Echo State Networks, Edge Detection, Feature Vector, Octonion, Octonion Vector Product

References

[1] EVANS C J，SANGWINE S J， ELL T A: Proceedings of the Tenth European Signal Processing Conference． Vol. 1 (2000): pp.107-110.

[2] LwataH，AguiT，Nagahashi H: IEEE International Conference on Neural Networks. Vol. 3 (1999): pp.1426-1431.

[3] Weiqing Li, Chengbiao Wang, Qun Wang, et al: International Symposium on Information Science and Engineering. Vol. 2 (2008): pp.40-44.

DOI: 10.1109/isise.2008.310

[4] Jaeger H，Lukosevicius M，Popovici D，Udo Siewert: Neural Networks. Vol. 20 (2007): pp.335-352.

[5] J WANG, L LIU: Proceedings of Intermational Congress on Image and Signal Processing. Vol. 1 (2010): pp.1137-1141.

[6] JIANGYAN XU, LINNING YE, WANG LUO: International Journal of Imaging Systems and Technology. Vol. 20, No. 4 (2010): pp.354-358.