目标检测与识别:视觉显著性与似物性采样

显著图是显示每个像素独特性的图像。显著图的目标在于将一般图像的表示简化或是改变为更容易分析的样式。本章讲解了什么是视觉显著性,提取显著性区域包括Buttom-up方法、Top-down方法两大类,似物性采样研究现状及方法等内容。
展开查看详情

1.第十二章 目标检测与识别 Lecture 12 Object Detection and Recognition

2.12.1 视觉显著性与似物性采样

3.内容 什么是视觉显著性 怎样提取显著性区域 似物性采样

4.什么是显著图?

5.

6.

7.What is Saliency map? The Saliency Map is a topographically arranged map that represents visual saliency of a corresponding visual scene.

8.How to detect Saliency map? Butto m -up approach L. Itti’s approach Spectral Residual approach Frequency-tuned approach Global contrast based approach Top-down approach Context-aware

9.How to detect Saliency map? Button-up approach L. Itti’s approach Spectral Residual approach Frequency-tuned approach Global contrast based approach Top-down approach Context-aware

10.How to detect Saliency map? Button-up approach L. Itti’s approach Spectral Residual approach Frequency-tuned approach Global contrast based approach Top-down approach Context-aware

11.L. Itti’s approach (TPAMI,1998) Gaussian Pyramids R , G , B , Y Gabor pyramids for q = {0º, 45º, 90º, 135º}

12.Gaussian Pyramid

13.Gabor Filter Gabor filter , is a linear filter used for edge detection, texture representation and discrimination. In the spatial domain, a 2D Gabor filter is a Gaussian kernel function modulated by a sinusoidal plane wave. J. G. Daugman discovered that simple cells in the visual cortex of mammalian brains can be modeled by Gabor functions. Thus, image analysis by the Gabor functions is similar to perception in the human visual system.

14.Gabor Filter

15.Gabor Filter

16.视觉神经结构 感受野 : 直接或间接影响某一特定神经细胞的光感受器细胞的全体

17.同心圆感受野 人的视觉细胞存在视觉场结构 . 视点的中心区域存在正性细胞.它们接收光能并产生一个正的反应。在该中心区域周围存在着负性细胞.它们在接收光能时产生相反的反应。负性细胞随中心距增大而迅速稀疏,代之而起的中性细胞不产生任何反应。这种解释由诺贝尔奖金获得者 Hartline 得到证实。 这种场结构所产生的视觉反应可由 ” 墨西哥草帽 ” 来表示 . 这种场结构可以使人的视觉具有侧抑制作用,它使观察物体时保证 “ 集中注意力 ” .即把视觉活动集中在注意圈内,不受圈外的变化所干扰。

18.同心圆感受野

19.感受野同心圆拮抗式模型 ( Rodieck , 1965 )

20.同心圆感受野工作原理

21. 马赫带效应

22.马赫带-指人们在明暗变化的边界,常常在亮区看到一条更亮的光带,而在暗区看到一条更暗的线条。这就是马赫带现象,马赫带不是由于刺激能量的分布,而是由于受到视觉”惰性”的影响

23.马赫带-指人们在明暗变化的边界,常常在亮区看到一条更亮的光带,而在暗区看到一条更暗的线条。这就是马赫带现象,马赫带不是由于刺激能量的分布,而是由于受到视觉”惰性”的影响

24.Laplacian of Gaussian (LOG)  

25.Relations between DOG and LOG Various derivatives are: Therefore,  

26.L. Itti’s approach (TPAMI,1998) Center-surround Difference Achieve center-surround difference through across-scale difference Operated denoted by Q: Interpolation to finer scale and point-to-point subtraction One pyramid for each channel: I( s ), R( s ), G( s ), B( s ), Y( s ) where s Î [0..8] is the scale

27.L. Itti’s approach (TPAMI,1998) Center-surround Difference Intensity Feature Maps I ( c , s ) = | I( c ) Q I( s ) | c Î {2, 3, 4} s = c + d where d Î {3, 4} So I (2, 5) = | I (2) Q I (5)| I (2, 6) = | I (2) Q I (6)| I (3, 6) = | I (3) Q I (6)| …  6 Feature Maps

28.L. Itti’s approach (TPAMI,1998) Center-surround Difference Color Feature Maps Red-Green and Yellow-Blue Center-surround Difference Orientation Feature Maps +R-G +R-G +G-R +G-R +B-Y +Y-B +Y-B +B-Y +B-Y Same c and s as with intensity RG ( c , s ) = | (R( c ) - G( c )) Q (G( s ) - R( s )) | BY ( c , s ) = | (B( c ) - Y( c )) Q (Y( s ) - B( s )) |

29.L. Itti’s approach (TPAMI,1998) Normalization Operator Promotes maps with few strong peaks Surpresses maps with many comparable peaks