中国组织工程研究 ›› 2010, Vol. 14 ›› Issue (30): 5568-5572.doi: 10.3969/j.issn.1673-8225.2010.30.015

• 骨与关节图像与影像 bone and joint imaging • 上一篇    下一篇

基于稀疏性与同步性的视皮质假体图像处理策略

师  黎,刘旭辉,钟丽辉   

  1. 郑州大学电气工程学院,河南省郑州市  450001
  • 出版日期:2010-07-23 发布日期:2010-07-23
  • 通讯作者: 刘旭辉,硕士,郑州大学电气工程学院,河南省郑州市 450001 liuxuhui0119@163.com
  • 作者简介:师 黎☆,女,1964年生,河南省尉氏县人,汉族,2007年上海大学毕业,博士,教授,博士生导师,主要研究领域为智能控制、生物信号分析处理。 shili@zzu.edu.cn
  • 基金资助:

    国家自然科学基金(60841004):基于基函数超完备集的动物视觉图像重构研究;国家自然科学基金(60971110):初级视觉皮层中视像整体特征的稀疏表象模型的研究。

Visual image processing strategy based on sparsity and synchronization in the visual cortex prosthesis

Shi Li, Liu Xu-hui, Zhong Li-hui   

  1. School of Electrical Engineering, Zhengzhou University, Zhengzhou  450001, Henan Province, China
  • Online:2010-07-23 Published:2010-07-23
  • Contact: Liu Xu-hui, Master, School of Electrical Engineering, Zhengzhou University, Zhengzhou 450001, Henan Province, China liuxuhui0119@163.com
  • About author:Shi Li☆, Doctor, Professor, Doctoral supervisor, School of Electrical Engineering, Zhengzhou University, Zhengzhou 450001, Henan Province, China shili@zzu.edu.cn
  • Supported by:

    the National Natural Science Foundation of China, No. 60841004*, 60971110*

摘要:

背景:视觉系统是人类获知外部世界的重要感觉系统,因各种疾病或其他损伤导致视觉丧失的盲人数量与日俱增,视皮质假体的研究有望为盲人提供一条复明的途径。视皮质假体虽然经过多年的研究,在某些方面取得了一定的进展和成果,但仍存在很多问题和挑战。
目的:针对置入视皮质的刺激电极数量有限这一问题,提出一种基于稀疏性和同步性的视皮质假体图像处理策略。
方法:该策略借鉴视觉皮质的生理特性,首先采用基于稀疏性的拓扑独立成分分析方法提取自然图像的基函数及神经元的响应,然后采用基于同步性的脉冲耦合神经网络方法对响应系数进行筛选,获取最佳神经元响应,最后利用最佳神经元响应刺激相应的神经元达到恢复自然图像的目的。
结果与结论:实验结果表明,该策略理论上可以利用较少的神经元表征自然图像中的重要信息,应用于视皮质假体可有效减少置入刺激电极的数量,达到更好的信息传递效果。

关键词: 视皮质假体, 图像处理, 基函数, 拓扑独立分量分析, 脉冲耦合神经网络

Abstract:

BACKGROUND: Visual system is the effective and important sensory system for animals receiving outside information. The number of the blind patients who are caused by the visual diseases or visual injury is boosting, and the visual cortex prosthesis is a promising way for the patients to re-achieve the vision. Despite some progresses after several years research of the visual cortex prosthesis, there remain issues and challenges in visual cortex prosthesis.
OBJECTIVE: Because the number of the microelectrode which is implanted in the visual cortex is restricted, the visual image processing strategy based on sparsity and synchronization in the visual cortex prosthesis is proposed.
METHODS: The strategy was based on the physiological properties of the visual cortex. Firstly, the image basic functions were abstracted using the sparsity of the Topographic Independent Component Analysis (TICA). Then, the optimal responses of basic functions were selected using pulse-coupled neural networks (PCNN). Finally, the optimal responses were used to stimulate the corresponding neurons to restore the visual image.
RESULTS AND CONCLUSION: Simulated results demonstrated that the nature images were representative using less available neurons. Therefore, it could lessen number of the electrodes of the visual cortex prosthesis which were implanted in visual cortex for the blind people.

中图分类号: