Chinese

2019,  1 (2):   240 - 249   Published Date：2019-4-20

DOI: 10.3724/SP.J.2096-5796.2018.0013

Abstract

For the blind, it is difficult to appreciate calligraphy. A tactile generation technology of text image is developed, which can enable the blind to recognize text in digital images by tactile recognition. This paper presents a Gauss difference algorithm based on tangential flow of edges, which can extract the contour of text precisely, and get the smooth contour of text with remarkable edge features. A tactile generation algorithm based on five elements is proposed to generate stable and continuous tactile sense for perceiving the contour of text. The experimental results show that through some adaptive training, the blind can better recognize some characters with simple contours, which provides a possibility for tactile appreciation of calligraphy.

Content

1 引言

2 基于高斯差分的文字轮廓边缘提取

2.1　梯度图像的获取

$Y U V = 0.2990 0.5870 0.1440 - 0.1678 - 0.3313 0.5000 0.5000 - 0.4187 - 0.0813 R G B$

$Y = 0.299 R + 0.587 G + 0.144 B$

$Δ Y = Δ Y x Δ Y y = Y ( x - 1 , y + 1 ) + 2 Y ( x , y + 1 ) + Y ( x + 1 , y + 1 ) - Y ( x - 1 , y - 1 ) - 2 Y ( x , y - 1 ) - Y ( x + 1 , y - 1 ) Y ( x - 1 , y - 1 ) + 2 Y ( x - 1 , y ) + Y ( x - 1 , y + 1 ) - Y ( x + 1 , y - 1 ) - 2 Y ( x + 1 , y ) - Y ( x + 1 , y + 1 )$

$Δ Y = Δ Y x 2 + Δ Y y 2$

$Δ Y = Δ Y x + Δ Y y$

$X Y Z = 0.4124 0.3576 0.1805 0.2126 0.7152 0.0722 0.0193 0.1192 0.9505 R G B$
${L}^{\mathrm{*}}{a}^{\mathrm{*}}{b}^{\mathrm{*}}$ 的值即为：
$L * = 116 Y / Y 0 1 / 3 - 16 a * = 500 X / X 0 1 / 3 - Y / Y 0 1 / 3 b * = 200 Y / Y 0 1 / 3 - Z / Z 0 1 / 3$

$Δ C = Δ L x * Δ a x * Δ b x * Δ L y * Δ a y * Δ b y * = L * ( x + 1 , y ) - L * ( x , y ) , a * ( x + 1 , y ) - a * ( x , y ) , b * ( x + 1 , y ) - b * ( x , y ) L * ( x , y + 1 ) - L * ( x , y ) , a * ( x , y + 1 ) - a * ( x , y ) , b * ( x , y + 1 ) - b * ( x , y )$

$Δ C = Δ E x * 2 + Δ E y * 2$

$Δ C = Δ E x * + Δ E y *$

$Δ E x * = Δ L x * 2 + Δ a x * 2 + Δ b x * 2 Δ E y * = Δ L y * 2 + Δ a y * 2 + Δ b y * 2$

$Δ E x * = Δ L x * + Δ a x * + Δ b x * Δ E y * = Δ L y * + Δ a y * + Δ b y *$

$Δ Y = ( Δ Y - m i n ) ( m a x - m i n )$
$Δ C = ( Δ C - m i n ) ( m a x - m i n )$

$G ( x , y ) = Δ Y ' 2 + Δ C ' 2$
2.2　边缘切线流的构造

ETF滤波器被定义为：
$t n e w ( x ) = 1 K ∑ y ∈ Ω ( x ) Φ ( x , y ) t c u r ( y ) w s ( x , y ) w m ( x , y ) w d ( x , y )$

$w s ( x , y ) = 1 , 若 X - Y < r 0 , 其 他 情 况$

$w m ( x , y ) = 1 2 ( 1 + t a n h η ⋅ ( e ( y ) - e ( x ) ) )$

$w d ( x , y ) = t c u r ( x ) ⋅ t c u r ( y )$

$Φ ( x , y ) = 1 , 若 t c u r ( x ) ⋅ t c u r ( y ) > 0 - 1 , 其 他 情 况$

2.3　轮廓边缘滤波

$F ( s ) = ∫ - T T I ( l s ( t ) ) f ( t ) d t$

$f ( t ) = G σ c ( t ) - ρ ⋅ G σ s ( t )$

$G σ ( x ) = 1 2 π σ e - x 2 2 σ 2$

$H ( x ) = ∫ - S S G σ m ( s ) F ( s ) d s$

$H ˜ ( x ) = 0 , H ( x ) < 0 或 1 + t a n h ( H ( x ) ) < τ 1 , 其 他 情 况$

2.4　文字轮廓边缘提取的实验结果对比

3 基于五要素的触觉生成算法

3.1　碰撞检测

3.2　触力觉呈现模型

HIP沿着物体表面(由三角面片构成)滑移时，并不是在平面上的简单直线或曲线运动，而是高低不平的三角面片间的移动。为了反馈力计算的精确性，将三角面片的法矢方向作为God-Object的刺穿深度方向，利用胡克定理对三角面片计算法向力。如图6所示，假设 ${q}_{1}$ 为HIP在前一时刻的空间位置， ${q}_{2}$ 为HIP在后一时刻的空间位置，HIP从 ${q}_{1}$ 移动到 ${q}_{2}$ 过程中，与三维纹理图形中的三角面片ABC在 $x$ 处发生碰撞，根据God-Object模型，可以得到后一时刻God-Object位置 $Q$ ，那么 ${q}_{2}$$Q$ 点之间的距离 $\stackrel{\to }{h}$ 即为HIP刺透模型的深度，方向由三角面片ABC的法向量 $\stackrel{\to }{{F}_{n0}}$ 方向决定， $k$ 为刚性系数。那么法向力就可以计算为：
$F n 0 → = k h →$

$F c → = c V →$

$F μ → = - μ F n 0 →$

$F g → = F n 0 → + F c → + F μ →$
4 系统测试与评价

4.1　测试系统平台搭建

4.2　实验与分析

4 结语

Reference

1.

Liu Y J, Yu M J, Fu Q F, Chen W F, Liu Y, Xie L X. Cognitive mechanism related to line drawings and its applications in intelligent process of visual media: A survey. Frontiers of Computer Science, 2016, 10(2): 216–232 DOI:10.1007/s11704-015-4450-1

2.

Zhang S. Research and application of edge detection algorithms based on machine vision. Shanghai: Shanghai Jiaotong University, 2013

3.

da Silva R D, Minetto R, Schwartz W R, Pedrini H. Adaptive edge-preserving image denoising using wavelet transforms. Pattern Analysis and Applications, 2013, 16(4): 567–580 DOI:10.1007/s10044-012-0266-x

4.

Chai M, Shao T, Wu H, Weng Y, Zhou K. AutoHair: fully automatic hair modeling from a single image. Acm Transactions on Graphics. 2016, 35(4): 1–12 DOI:10.1145/2897824.2925961

5.

Zhu X, Chen M, Li X, Zhu Y. A color image edge detection model with morphological amoebas and two kinds of color spaces. Journal of Computer-Aided Design & Computer Graphics, 2014, 26(7): 1060–1066

6.

Quan W, Huang H. Contour extraction of multi-feature orientation preference. Journal of Computer-Aided Design & Computer Graphics, 2018, 30(1): 100–106 DOI:10.3724/SP.J.1089.2018.16227

7.

Tang Q, Sang N, Liu H, Chen X. Detecting natural image contours by combining visual perception and machine learning. 2013, 43(9): 1124–1135 DOI:10.1360/112012-630

8.

Knott T C, Kuhlen T W. Accurate and adaptive contact modeling for multi-rate multi-point haptic rendering of static and deformable environments. Computers & Graphics, 2016, 57: 68–80 DOI:10.1016/j.cag.2016.03.007

9.

Stanley A A, Okamura A M. Deformable model-based methods for shape control of a haptic jamming surface. IEEE Transactions on Visualization and Computer Graphics, 2017, 23(2): 1029–1041 DOI:10.1109/tvcg.2016.2525788

10.

Dervaux F, Peterlik I, Dequidt J, Cotin S, Duriez C. Haptic rendering of interacting dynamic deformable objects simulated in real-time at different frequencies. In: 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems. 2013, 2010-2016 DOI:10.1109/IROS.2013.6696624

11.

Yang W, Gao S, Wan H, Zhu Z, Luo Y. Physically-based haptic display of rigid body collisions with virtual hand interaction. Chinese Journal of Computers, 2006, 29(12): 2096–2103 DOI:10.3321/j.issn:0254-4164.2006.12.005

12.

Wu J, Song A, Li J. Research on mass-spring force/deformation modeling for haptic display. Journal of system simulation, 2006, 18(11): 3152–3156 DOI:10.3969/j.issn.1004-731X.2006.11.037

13.

Zhang L, Wang Y, Wu X. GPU-based expressive line drawing algorithm for volumes. 2014, 44(11), 1385–1397 DOI:10.1360/N112014-00203

14.

Kovler I, Joskowicz L, Weil Y A, Khoury A, Kronman A, Mosheiff R, Liebergall M, Salavarrieta J. Haptic computer-assisted patient-specific preoperative planning for orthopedic fractures surgery. International Journal of Computer Assisted Radiology and Surgery, 2015, 10(10): 1535–1546 DOI:10.1007/s11548-015-1162-9