color image processing method and applications - lukac, plataniotis 2007

P1: Shashi September 14, 2006 15:28 9774 9774˙C000 P1: Shashi September 14, 2006 15:28 9774 9774˙C000 P1: Shashi September 14, 2006 15:28 9774 9774˙C000 P1: Shashi September 14, 2006 15:28 9774 9774˙C000 CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2007 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S Government works Printed in the United States of America on acid-free paper 10 International Standard Book Number-10: 0-8493-9774-X (Hardcover) International Standard Book Number-13: 978-0-8493-9774-5 (Hardcover) This book contains information obtained from authentic and highly regarded sources Reprinted material is quoted with permission, and sources are indicated A wide variety of references are listed Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers For permission to photocopy or use material electronically from this work, please access www.copyright.com (http:// www.copyright.com/) or contact the Copyright Clearance Center, Inc (CCC) 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400 CCC is a not-for-profit organization that provides licenses and registration for a variety of users For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com P1: Shashi September 14, 2006 15:28 9774 9774˙C000 Color television! Bah, I won’t believe it until I see it in black and white —Samuel Goldwyn, movie producer P1: Shashi September 14, 2006 15:28 9774 9774˙C000 P1: Shashi September 14, 2006 15:28 9774 9774˙C000 Dedication To my dear parents, whose constant love and support have made my achievements possible R Lukac To the loving memory of my father K.N Plataniotis P1: Shashi September 14, 2006 15:28 9774 9774˙C000 P1: Shashi September 14, 2006 15:28 9774 9774˙C000 Preface Over the last two decades, we have witnessed an explosive growth in both the diversity of techniques and the range of applications of image processing However, the area of color image processing is still sporadically covered, despite having become commonplace, with consumers choosing the convenience of color imaging over traditional grayscale imaging With advances in image sensors, digital TV, image databases, and video and multimedia systems, and with the proliferation of color printers, color image displays, DVD devices, and especially digital cameras and image-enabled consumer electronics, color image processing appears to have become the main focus of the image-processing research community Processing color images or, more generally, processing multichannel images, such as satellite images, color filter array images, microarray images, and color video sequences, is a nontrivial extension of the classical grayscale processing Indeed, the vectorial nature of multichannel images suggests a different approach — that of vector algebra and vector fields — should be utilized in approaching this research problem Recently, there have been many color image processing and analysis solutions, and many interesting results have been reported concerning filtering, enhancement, restoration, edge detection, analysis, compression, preservation, manipulation, and evaluation of color images The surge of emerging applications, such as single-sensor imaging, color-based multimedia, digital rights management, art, and biomedical applications, indicates that the demand for color imaging solutions will grow considerably in the next decade The purpose of this book is to fill the existing literature gap and comprehensively cover the system, processing and application aspects of digital color imaging Due to the rapid developments in specialized areas of color image processing, this book has the form of a contributed volume, in which well-known experts address specific research and application problems It presents the state-of-the-art as well as the most recent trends in color image processing and applications It serves the needs of different readers at different levels It can be used as a textbook in support of a graduate course in image processing or as a stand-alone reference for graduate students, researchers, and practitioners For example, the researcher can use it as an up-to-date reference, because it offers a broad survey of the relevant literature Finally, practicing engineers may find it useful in the design and the implementation of various image- and video-processing tasks In this book, recent advances in digital color imaging and multichannel image-processing methods are detailed, and emerging color image, video, multimedia, and biomedical processing applications are explored The first few chapters focus on color fundamentals, targeting three critical areas: color management, gamut mapping, and color constancy The remaining chapters explore color image processing approaches across a broad spectrum of emerging applications ranging from vector processing of color images, segmentation, resizing and compression, halftoning, secure imaging, feature detection and extraction, image retrieval, semantic processing, face detection, eye tracking, biomedical retina image analysis, real-time processing, digital camera image processing, spectral imaging, enhancement for plasma display panels, virtual restoration of artwork, image colorization, superresolution image reconstruction, video coding, video shot segmentation, and surveillance Discussed in Chapters to are the concepts and technology essential to ensure constant color appearance in different devices and media This part of the book covers issues related P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color (a) (b) (c) COLOR FIGURE 8.13 Color shares obtained using the nonexpansive ISS solution when cryptographic processing is performed for two color channels: (a) RG channels with c = and c = 2, (b) RB channels with c = and c = 3, and (c) GB channels with c = and c = (a) (b) (c) (d) COLOR FIGURE 9.8 In columns, respectively, (a) input image, (b) RGB-gradient-based saliency map, (c) color-boosted saliency map, and (d) the results with red dots (lines) for gradient-based method and yellow dots (lines) for salient points after color saliency boosting P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color COLOR FIGURE 10.14 A query by image example (left), and the corresponding retrieval set (right) COLOR FIGURE 10.15 A query by sketch (left), and the corresponding retrieval set (right) P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color (a) (b) COLOR FIGURE 10.19 (a) The target images used in the test; and (b) a user’s query sketches for the eight images Light from fluorescent lamps at the ceiling Daylight from the window COLOR FIGURE 12.4 The face is illuminated by the nonuniform illumination field, and the white balancing partially fails The color appearance of the face varies at different parts of the light field Horizon 2300 K Incandescent A 2856 K 0.45 TL 84 (F11) 4000 K Daylight 6500 K Canonical chromacities of skin tones NCC g 0.4 0.35 0.3 0.25 0.45 0.5 0.55 0.6 0.65 0.7 NCC r COLOR FIGURE 12.7 The skin tone appearance difference can be clearly observed in the four images taken with the Sony camera (see Figure 12.2) From the selected area marked with a box, the RGB values were taken and converted to NCC color space As shown in the graph below the images, the areas of canonical chromaticities more or less overlap P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color Color temperature of the prevailing illumination Color temperature of the prevailing illumination COLOR FIGURE 12.8 The color appearance shift The color temperature of the light sources increases from left to right The arrow indicates the change in the color of the light The limited dynamic response range causes distortion in color: pixels can saturate to a maximum value (the rightmost image at the upper row) or be underexposed to zero (the leftmost image at the lower row) (a) (b) (c) (d) COLOR FIGURE 13.8 Tracking the iris while staying outdoors and additionally changing the illumination conditions by altering between IR to non-IR lighting Notice how light conditions change when switching between IR and non-IR light emission (greenish looking images are IR “night vision”) P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color COLOR FIGURE 13.9 Tracking the iris of Asians under scale changes and heavy light disturbances (a) (b) (c) COLOR FIGURE 14.1 Normal and abnormal images: (a) a typical normal image with optic disc region, (b) severe retinopathy, with large plaques of exudates, and (c) diabetic retinopathy including retinal exudates and a cotton wool spot (a) (b) (c) COLOR FIGURE 14.7 Connected component labeling and boundary tracing algorithms results: (a) a typical retinal image, (b) FCM segmented image, and (c) labeled regions using connected component approach P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color COLOR FIGURE 16.1 Color image representation in the RGB color domain (a) (b) (c) (d) (e) (f ) (g) (h) COLOR FIGURE 16.6 Examples of RGB CFAs: (a) Bayer pattern, (b) Yamanaka pattern, (c) diagonal stripe pattern, (d) vertical stripe pattern, (e) diagonal Bayer pattern, (f,g) pseudo-random patterns, and (h) HVS-based pattern P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color (a) (b) COLOR FIGURE 16.9 CFA image obtained using a well-known Bayer CFA with the GRGR phase in the first row: (a) acquired grayscale CFA image and (b) CFA image rearranged as a color image (a) (b) (c) COLOR FIGURE 16.16 Performance improvements obtained by changing the SM: (a) original image, (b) Kimmel algorithm based on the color-ratio model (Kimmel, R., IEEE Trans on Image Process., 8, 1221, 1999), and (c) Kimmel algorithm based on the normalized color-ratio model (Lukac, R., and Plataniotis, K.N., IEEE Trans on Consumer Electron., 50, 737, 2004) P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color (a) (b) COLOR FIGURE 16.17 Demosaicking process: (a) restoration of the image shown in Figure 16.9a using the demosaicking solution and (b) image shown in Figure 16.17a enhanced using the demosaicked image postprocessor (a) (b) (c) (d) COLOR FIGURE 16.18 Influence of the ESM and SM on the quality of the demosaicked image: (a) used both ESM and SM, (b) omitted ESM, (c) omitted SM, and (d) omitted both ESM and SM P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color (a) (b) (c) (d) (e) (f ) (g) (h) (i) COLOR FIGURE 16.19 Influence of the CFA on the quality of the demosaicked image demonstrated using the same processing solution: (a–h) demosaicked image respectively corresponding to the RGB CFAs shown in Figure 16.6a to Figure 16.6h and (i) original image (a) (b) (c) (a) (b) (c) COLOR FIGURE 16.24 Median filtering based spatial interpolation (R Lukac, K.N Plataniotis, B Smolka, and A.N Venetsanopulos, in Proceedings of the IEEE International Symposium on Industrial Electronics, III, 1273, 2005) The cropped patterns correspond to: (a) original (small) images, and (b, c) upsampled images obtained using (b) component-wise processing and (c) vector processing P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color COLOR FIGURE 17.12 Classification of polymer waste: (left) input image and (right) after classification The classification image shows the assignment of each pixel to the material PP (injection blow molding [IBM] quality), PETP, PP (deep drawing [DD] quality), S/B (styrenebutadiene copolymer), and PE-HD COLOR FIGURE 17.15 Binomial prediction of tumor probability in a one-dimensional subspace resulting from a dimension reduction step (compare densities of Figure 17.14) The predicted tumor probabilities can be mapped out using the color bar shown on the right, see Figure 17.16 (a) (b) (c) (d) COLOR FIGURE 17.16 Color maps indicating the tumor probability computed automatically from MR spectral images (a, c), superimposed onto morphologic MR images (b, d) The color mapping used is defined on the right-hand side of Figure 17.15 P1: Naresh Chandra August 23, 2006 13:46 9774 (a) 9774˙Color (b) (c) (d) COLOR FIGURE 18.5 Simulation of dynamic false contours: (a) original image, (b) simulation with [1 16 32 48 48 48 48], (c) simulation with [1 16 32 42 44 52 54], and (d) simulation with error diffusion (a) (b) (c) (d) COLOR FIGURE 18.13 Constant image with 8-bit coded fraction = 32 by error diffusion-based technique: (a, b) result of conventional error diffusion and its enlarged version, and (c, d) result of described technique and its enlarged version (a) (b) (c) (d) COLOR FIGURE 18.18 Constant image with 4-bit coded fraction = by dithering-based technique: (a, b) result of conventional error diffusion and its enlarged version, and (c, d) result of described technique and its enlarged version P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color COLOR FIGURE 18.20 Gamut difference between PDP and CRT COLOR FIGURE 20.1 Still image colorization examples Given a grayscale image (left), the user marks chrominance scribbles (center), and our algorithm provides a colorized image (right) The image size/run-times top to bottom are 270 × 359/less than 0.83 seconds, 256 × 256/less than 0.36 seconds, and 400 × 300/less than 0.77 seconds P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color (a) (b) (c) COLOR FIGURE 20.6 Recolorization example (a) Original color image, (b) shows the scribbles placed by the user, where white is a special scribble used in recolorization that preserves the original colors of the image in selected regions of the image, and (c) is the recolored image Note that only a few scribbles were used Applying more scribbles would increase the quality of the result image (a) (b) (c) COLOR FIGURE 20.7 Recolorization example using the Cr channel for measurement (M) rather than the intensity channel (a) is the original color image, and (b) is the intensity channel of the image It can be clearly seen that the intensity image does not contain significant structural information on the red rainbow strip (this is a unique example of this effect) On the other hand, both the Cb and Cr change significantly between the stripes and therefore can be used for recolorization (c) Recolored image in which the red stripe of the rainbow was replaced by a purple one (a) (b) COLOR FIGURE 20.9 Using the described technique to remove color from certain image regions: (a) original image, and (b) only the eyes from the original image are preserved P1: Naresh Chandra August 23, 2006 13:46 9774 9774˙Color (a) (b) (c) (d) COLOR FIGURE 23.2 Examples of gradual transitions: (a) dissolve, (b) fade, (c) a classical wipe, and (d) a wipe of the “curtain” variety (a) (b) COLOR FIGURE 24.7 Example of color-based skin detection/segmentation: (a) original image, and (b) image after skin detection ... design and the implementation of various image- and video -processing tasks In this book, recent advances in digital color imaging and multichannel image- processing methods are detailed, and emerging... Real-Time Imaging, Special Issue on Multi-Dimensional Image Processing, and of the Computer Vision and Image Understanding, Special Issue on Color Image Processing for Computer Vision and Image. .. United States of America on acid-free paper 10 International Standard Book Number-10: 0-8 49 3-9 774-X (Hardcover) International Standard Book Number-13: 97 8-0 -8 49 3-9 77 4-5 (Hardcover) This book contains