Emergence of simple-cell receptive field properties by learning a sparse code for natural images
Emergence of simple-cell receptive field properties by learning a sparse code for natural images
复制标题
DOI：
10.1038/381607a0
复制
发表时间：
1996-06-13
期刊：
影响因子：
64.8
通讯作者：
Field, DJ
中科院分区：
文献类型：
--
作者：
Olshausen, BA;Field, DJ
哺乳动物初级视觉皮层中简单细胞的感受野可以被表征为空间定位、定向（1-4）和带通（对不同空间尺度的结构具有选择性），与小波变换的基函数相比（5,6），理解视觉神经元这种响应特性的一种方法是从有效编码的角度考虑它们与自然图像的统计结构的关系（7-12），沿着这些思路， 许多研究尝试在自然图像上训练无监督学习算法，希望开发具有类似属性的感受野（13-18），但没有一个研究成功地产生跨越图像空间并包含上述所有三个属性的完整集合。在这里，我们研究了建议（8,12），即最大化稀疏性的编码策略足以解释这些属性，我们表明，尝试为自然场景找到稀疏线性代码的学习算法将开发一整套局部的、定向的、带通感受野，类似于在初级视觉皮层中发现的感受野。由此产生的稀疏图像代码为后期处理阶段提供了更有效的表示，因为它具有更高的 其产出之间的统计独立性程度。
THE receptive fields of simple cells in mammalian primary visual cortex can be characterized as being spatially localized, oriented(1-4) and bandpass (selective to structure at different spatial scales), comparable to the basis functions of wavelet transforms(5,6), One approach to understanding such response properties of visual neurons has been to consider their relationship to the statistical structure of natural images in terms of efficient coding(7-12), Along these lines, a number of studies have attempted to train unsupervised learning algorithms on natural images in the hope of developing receptive fields with similar properties(13-18), but none has succeeded in producing a full set that spans the image space and contains all three of the above properties. Here we investigate the proposal(8,12) that a coding strategy that maximizes sparseness is sufficient to account for these properties, We show that a learning algorithm that attempts to find sparse linear codes for natural scenes will develop a complete family of localized, oriented, bandpass receptive fields, similar to those found in the primary visual cortex, The resulting sparse image code provides a more efficient representation for later stages of processing because it possesses a higher degree of statistical independence among its outputs.