10 April 2018 Feature extraction inspired by V1 in visual cortex
Author Affiliations +
Proceedings Volume 10615, Ninth International Conference on Graphic and Image Processing (ICGIP 2017); 106155C (2018) https://doi.org/10.1117/12.2302951
Event: Ninth International Conference on Graphic and Image Processing, 2017, Qingdao, China
Target feature extraction plays an important role in pattern recognition. It is the most complicated activity in the brain mechanism of biological vision. Inspired by high properties of primary visual cortex (V1) in extracting dynamic and static features, a visual perception model was raised. Firstly, 28 spatial-temporal filters with different orientations, half-squaring operation and divisive normalization were adopted to obtain the responses of V1 simple cells; then, an adjustable parameter was added to the output weight so that the response of complex cells was got. Experimental results indicate that the proposed V1 model can perceive motion information well. Besides, it has a good edge detection capability. The model inspired by V1 has good performance in feature extraction and effectively combines brain-inspired intelligence with computer vision.
© (2018) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Chao Lv, Yuelei Xu, Xulei Zhang, Shiping Ma, Shuai Li, Peng Xin, Mingning Zhu, Hongqiang Ma, "Feature extraction inspired by V1 in visual cortex", Proc. SPIE 10615, Ninth International Conference on Graphic and Image Processing (ICGIP 2017), 106155C (10 April 2018); doi: 10.1117/12.2302951; https://doi.org/10.1117/12.2302951


Fighting detection using interaction energy force
Proceedings of SPIE (February 08 2017)
Weakly supervised specific object modelling for recognition
Proceedings of SPIE (October 30 2009)
Bilinear models of natural images
Proceedings of SPIE (February 07 2007)
Physically based and probabilistic models for computer vision
Proceedings of SPIE (September 01 1991)

Back to Top