2
Abstract--This paper presents an image processing technique for assisting hypochromatopsia people to view the natural scenes. We propose an image enhancement approach based on the color space calculation and linear transform to adjust the color presentation, which is also called color reprint. The method can help hypochromatopsia people being able to distinguish the difference between the colors they usually confuse. The experimental results show that the proposed image enhancement technique is useful for hypochromatopsia people to recognize, separate the colors in natural scenes. I. INTRODUCTION This article addresses the color reprint techniques to assist the people with color vision deficiency (See Fig. 1). The people so-called color blindness or hypochromatopsia can only view a part of the visible spectrum (approximately 400 nm – 700 nm) compared to those with normal vision capability. The color blindness or hypochromatopsia are usually caused by two major reasons: natural genetic factors, and impaired nerve or brain. Most literatures show that the people with color blindness are about 10% of the world population, and currently there are no solutions to cure in both science and medicine fields. II. COLOR CHROMA REPRINTING APPROACH In this section, we present our approach for color blindness enhancement. Some confused colors in a scene can be distinguished using the proposed reprinting method for hypochromatopsia, and a flowchart is illustrated in Fig. 2. A. Color Space Computing We suggest using the reprinting matrices to shift the image color pixel values from a normal view to three better views for protanomaly, deuteranomaly and tritanomaly. It is based on the RGB color representation and the computed ΔL, ΔC and ΔH images via CIELAB transform for the confused image pixels. Suppose ΔL, ΔC and ΔH are three image channels and (R1, G1, B1) and (R2, G2, B2) are the colors of two corresponding image pixels represented in the RGB color space. We apply them to CIELAB for color computing, where ΔE is a pixel-based Euclidean distance, and k C , k H are both unity. Geometrically, the quantity H corresponds to the arithmetic mean of the chord lengths of the equal chroma circles of the two colors [17]. The parameters L = R 1 -R 2 and C, H are defined as follows, To reprint the better viewing images for hypochromatopsia, we adopt the linear transform formula and applied it to the CLCH image, the enhanced image I E is then computed as follows An Enhancement Technique for Hypochromatopsia Assisted Vision Huei-Yung Lin 1 , Li-Qi Chen 1 , and Min-Liang Wang 2 1 Department of Electrical Engineering, National Chung Cheng University, Taiwan 2 Asian Institute of TeleSurgery / IRCAD-Taiwan (a) The normal and the proposed enhanced images for protanomaly, deuteranomaly and tritanomaly, respectively (b) Simulated viewing images of hypochromatopsia in normal, protanomaly, deuteranomaly and tritanomaly, respectively [2]. Fig 1. (a) The normal and our enhanced images. (b) The simulated viewing images of who with hypochromatopsia. The images show that our enhanced method improves the results for hypochromatopsia to recognize the check patterns. Fig 2. The flowchart of the proposed color reprinting method. 978-1-4673-6199-6/13/$31.00 ©2013 IEEE 2013 IEEE 17th International Symposium on Consumer Electronics (ISCE) 225

[IEEE 2013 IEEE 17th International Symposium on Consumer Electronics (ISCE) - Hsinchu City, Taiwan (2013.06.3-2013.06.6)] 2013 IEEE International Symposium on Consumer Electronics

Embed Size (px)

Citation preview

Page 1: [IEEE 2013 IEEE 17th International Symposium on Consumer Electronics (ISCE) - Hsinchu City, Taiwan (2013.06.3-2013.06.6)] 2013 IEEE International Symposium on Consumer Electronics

Abstract--This paper presents an image processing technique

for assisting hypochromatopsia people to view the natural scenes. We propose an image enhancement approach based on the color space calculation and linear transform to adjust the color presentation, which is also called color reprint. The method can help hypochromatopsia people being able to distinguish the difference between the colors they usually confuse. The experimental results show that the proposed image enhancement technique is useful for hypochromatopsia people to recognize, separate the colors in natural scenes.

I. INTRODUCTION This article addresses the color reprint techniques to assist

the people with color vision deficiency (See Fig. 1). The people so-called color blindness or hypochromatopsia can only view a part of the visible spectrum (approximately 400 nm – 700 nm) compared to those with normal vision capability. The color blindness or hypochromatopsia are usually caused by two major reasons: natural genetic factors, and impaired nerve or brain. Most literatures show that the people with color blindness are about 10% of the world population, and currently there are no solutions to cure in both science and medicine fields.

II. COLOR CHROMA REPRINTING APPROACH In this section, we present our approach for color blindness

enhancement. Some confused colors in a scene can be distinguished using the proposed reprinting method for hypochromatopsia, and a flowchart is illustrated in Fig. 2.

A. Color Space Computing We suggest using the reprinting matrices to shift the image

color pixel values from a normal view to three better views for protanomaly, deuteranomaly and tritanomaly. It is based on the RGB color representation and the computed ΔL, ΔC and ΔH images via CIELAB transform for the confused image pixels. Suppose ΔL, ΔC and ΔH are three image channels and (R1, G1, B1) and (R2, G2, B2) are the colors of two corresponding image pixels represented in the RGB color space. We apply them to CIELAB for color computing,

where ΔE is a pixel-based Euclidean distance, and kC, kH are both unity. Geometrically, the quantity �H corresponds to the

arithmetic mean of the chord lengths of the equal chroma

circles of the two colors [17]. The parameters �L = R1-R2 and �C, �H are defined as follows,

To reprint the better viewing images for hypochromatopsia,

we adopt the linear transform formula and applied it to the CLCH image, the enhanced image IE is then computed as follows

An Enhancement Technique for Hypochromatopsia Assisted Vision Huei-Yung Lin1, Li-Qi Chen1, and Min-Liang Wang2

1Department of Electrical Engineering, National Chung Cheng University, Taiwan 2Asian Institute of TeleSurgery / IRCAD-Taiwan

(a) The normal and the proposed enhanced images for protanomaly,

deuteranomaly and tritanomaly, respectively

(b) Simulated viewing images of hypochromatopsia in normal,

protanomaly, deuteranomaly and tritanomaly, respectively [2]. Fig 1. (a) The normal and our enhanced images. (b) The simulated viewing images of who with hypochromatopsia. The images show that our enhanced method improves the results for hypochromatopsia to recognize the check patterns.

Fig 2. The flowchart of the proposed color reprinting method.

978-1-4673-6199-6/13/$31.00 ©2013 IEEE

2013 IEEE 17th International Symposium on Consumer Electronics (ISCE)

225

Page 2: [IEEE 2013 IEEE 17th International Symposium on Consumer Electronics (ISCE) - Hsinchu City, Taiwan (2013.06.3-2013.06.6)] 2013 IEEE International Symposium on Consumer Electronics

where Ii is an input normal image and the color sources CLCH data sequences are the reshaped forms. The color reprinting matrix MRC is an R3x3 matrix.

III. EXPERIMENTAL RESULTS In this paper, the linear color transform is adopted to

enhance the natural images for hypochromatopsia. It is based on the RGB color representation and �L, ΔC and ΔH images via CIELAB transform for computing the confused image pixels. The results show our transform of the RGB colors are highly effective for separating the confused colors. For example, the pixels of green leaves are changed to lower values and presented in ΔC and ΔH channels, and the rest pixels are presented in ΔL channel. It is then easily to process the color linear transform by the proposed reprinting method.

Figure 3 shows the results of color reprinting for a natural scene, and it is suitable for protanomaly, deuteranomaly and tritanomaly, respectively. The Vischeck [2] is used for simulating the images observed from the people with hypochromatopsia. In Fig. 3(b), the deuteranomaly people might be a bit confused for recognizing the flowers (red) and leaves (green). In contrast to Fig. 3(b), the proposed enhanced image shown in Fig. 3(d) is with less confusion.

IV. CONCLUSIONS This paper proposed an efficient color reprinting method to

assist hypochromatopsia for color perception. Several natural images are used to show the improvement of the proposed enhancement techniques. The proposed method might be useful for hypochromatopsia to recognize, separate colors of natural scenes. The proposed reprinting method will be evaluated by those with hypochromatopsia in future.

ACKNOWLEDGMENT The support of this work in part by the IRCADTaiwan of

Taiwan, R.O.C. under Grant RD-101012 and IRB-1010206 is gratefully acknowledged.

EXAMPLES OF REFERENCE STYLES [1] Nathans Thomas, Darcy Hogness, David S.: “Molec-ular Genetics of

Human Color Vision: The Genes En coding Blue, Green, and Red Pigments,” Science, New Series, Vol.232, No.4747, pp.193-202, 1986-04-11.

[2] Vischeck Web site: http://www.vischeck.com/ [3] CIE Web site: http://cie.co.at/ [4] Wyszecki, Gnther Stiles, W.S.: Color Science: Concepts and Methods,

Quantitative Data and Formulae (2nd ed.), 1982. [5] R. W. G. Hunt: The Reproduction of Colour (6th ed.), pp.112, 2004. [6] Carlson, Neil R.: Psychology: The Science of Behaviour, p.145, 2007. [7] Kalloniatis Michael and Luu Charles.: “Psychophysics of Vision: The

Perception of Color,” 2007-04-02. [8] Nathans J, Thomas D and Hogness DS: “Molecular genetics of human

color vision: the genes encoding blue, green, and red pigments,” Science, vol.232, no.4747,

[9] pp.193-202, 1986. [10] “Delta E: The Color Difference.”", Colorwiki.com, 2009-04-16 [11] Jia-Bin Huang, Yu-Cheng Tseng, Se-InWu and Sheng Jyh Wang:

“Information Preserving Color Transformation for Protanopia and Deuteranopia,” IEEE Signal Processing Letters, vol.14, no.10, pp.711-714, 2007-10.

[12] Jia-Bin Huang, Chu-Song Chen, Tzu-Cheng Jen, Sheng Jyh Wang.: “Image recolorization for the colorblind,” IEEE International Conference on Acoustics, Speech and Signal Processing, pp.1161-1164, 2009.

[13] Rasche K., Geist R. andWestall J.: “Detail preserving reproduction of color images for monochromats and dichromats,” IEEE Computer Graphics and Applications, vol.25, no.3, pp.22-30, 2005.

[14] Poret S., Dony R.D. and Gregori S.: “Image processing for colour blindness correction,” Science and Technology for Humanity (TIC-STH), 2009 IEEE Toronto International Conference, pp.539-544, 2009-09.

(a) The normal image.

(b) The simulated images for who has protanomaly, deuteranomaly and tritanomaly, respectively.

(c) The proposed enhanced images for protanomaly, deuteranomaly

and tritanomaly, respectively.

(d) The simulated color-blindness images by using the enhanced images (c) as input for protanomaly, deuteranomaly and tritanomaly, respectively [2].

Fig 3. The (a) are the normal and the enhanced images and (b) are the simulated images of who with hypochromatopsia. The images show that the enhanced

method is much better to recognize the color blindness check patterns.

2013 IEEE 17th International Symposium on Consumer Electronics (ISCE)

226