生物启发计算在图像与视频处理中的创新理论与应用

需积分: 9 176 浏览量更新于2024-07-19 收藏 12.41MB PDF 举报

《生物启发计算在图像与视频处理中的应用》是一本于2018年出版的专业书籍，由D.P. Acharjya和V. Santhi编著，ISBN-10: 1498765920，共436页，以PDF格式提供，大小为12MB。本书聚焦于生物启发计算在图像和视频处理领域的前沿理论和发展，特别关注了这些技术在众多实际应用中的重要性，如医疗成像安全、生物特征识别、遥感、纹理理解、模式识别以及基于内容的图像检索等。生物启发计算（Bio-Inspired Computing）是一种借鉴自然界生物系统解决问题策略的计算机科学分支，它将生物学原理与计算机算法相结合，旨在开发高效且具有智能的解决方案。在图像处理方面，该方法借鉴了诸如昆虫视觉系统（如飞蛾对光的敏感度和方向感知）、神经网络的结构（如模仿人脑处理视觉信息的方式）以及进化算法（如模拟自然选择优化图像处理参数）等特性。通过模拟生物系统的复杂性和适应性，生物启发计算可以提高图像分析的精度、鲁棒性和实时性。书中详细探讨了如何利用生物启发方法改进传统图像处理技术，例如，通过模拟视觉皮层的结构和功能来增强图像分割和特征提取；利用蜂群优化算法进行图像压缩，减少数据量而保持图像质量；或者运用模拟昆虫视觉机制的算法来改善目标检测和跟踪。对于视频处理，书中的内容可能涉及帧间插值、运动估计算法、视频编码和解码优化，以及实时视频分析中的生物启发行为模型。此外，书中还可能涵盖了生物启发计算在新兴领域如生物医学影像中的应用，比如用于疾病的早期诊断或治疗效果评估，以及在安全领域中，如利用生物特征识别技术如人脸识别、虹膜扫描或指纹识别来增强身份验证的安全性。在远程 sensing 领域，生物启发计算可能被用来解析卫星和无人机拍摄的高分辨率图像，帮助进行环境监测和地理空间数据分析。《生物启发计算在图像与视频处理中的应用》是一本深入浅出的教材，不仅阐述了理论概念，还提供了实用的案例研究和实践技巧，为图像和视频处理专业人士以及对该领域感兴趣的学生提供了宝贵的参考资料。通过这本书，读者不仅能了解到生物启发计算的最新进展，还能学习到如何将这些技术转化为实际问题的有效解决方案。

List of Figures xv

10.14 (a) Original, (b) noisy (Gaussian noise with σ = 0.25),

(f) DEML2 D-ﬁltered, (g) SORIGAML2D-ﬁltered Cameraman

image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286

11.1 Block diagram of the proposed work . . . . . . . . . . . . . 303

11.2 (3 × 3) pixel matrix . . . . . . . . . . . . . . . . . . . . . . . 304

11.3 Example of texture unit transformation . . . . . . . . . . . . 304

11.4 GWT of an input image . . . . . . . . . . . . . . . . . . . . 307

11.5 (3 × 3) convolution kernels for Prewitt operator . . . . . . . 309

11.6 Results of Prewitt operator-based edge detection . . . . . . 310

11.7 Results of region selection for Prewitt operator-based edges 310

11.8 K-NN illustration (k = 5) . . . . . . . . . . . . . . . . . . . 314

11.9 Accuracy of proposed system . . . . . . . . . . . . . . . . . 318

11.10 F-measure of prop osed system . . . . . . . . . . . . . . . . . 319

12.1 Typical object recognition system . . . . . . . . . . . . . . . 325

12.2 Columbia o bject image library (COIL − 100) . . . . . . . . 330

12.3 Training set of an object sampled at every 50

. . . . . . . . 331

12.4 Training set of an object sampled at every 25

. . . . . . . . 332

12.5 Eigenspace ma nifo lds for an object in the database with the

position angle manifold being sampled at 50

. . . . . . . . 332

12.6 Eigenspace ma nifo lds for an object in the database with the

position angle manifold being sampled at 25

. . . . . . . . 333

12.7 Comparision between PCA, ICA, SVD-PCA and per centage

recognition with n

nearest matching . . . . . . . . . . . . 334

12.8 (a) Comparision between PCA, ICA, and PCA-SVD for po-

sition angle sampling of 50

(b) Comparision between PCA,

ICA, and PCA-SVD for position angle sampling of 25

. . . 335

12.9 Comparision between PCA, ICA, and PCA-SVD for position

angle sampling of 10

. . . . . . . . . . . . . . . . . . . . . . 335

12.10 Results with po sition estimation . . . . . . . . . . . . . . . . 338

13.1 Block diagram of the proposed CBIR system . . . . . . . . . 346

13.2 Contrast detection ﬁlter showing inner and outer square

region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348

13.3 Sample input images from the COIL-100 da tabase . . . . . . 349

13.4 Sample input images from the Oxford ﬂower-102 database . 349

13.5 Sample segmented images of the COIL-100 DB shown in

Figure 13.3 . . . . . . . . . . . . . . . . . . . . . . . . . . . 350

13.6 Sample segmented images of the Oxford ﬂower-102 database

shown in Figure 13.4 . . . . . . . . . . . . . . . . . . . . . . 350

13.7 Orientation vector visualization of the COIL-100 data base

using HOG . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354

xvi List of Figures

13.8 Orientation vector visualization of the Oxford ﬂower-102

database using HOG . . . . . . . . . . . . . . . . . . . . . . 355

13.9 Magnitude vector visualization of the COIL -100 da tabase us-

ing HOG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356

13.10 Magnitude vector visualization of the Oxford ﬂower-102

database using HOG . . . . . . . . . . . . . . . . . . . . . . 357

13.11 Example for 2d-tree . . . . . . . . . . . . . . . . . . . . . . . 358

13.12 Insertion of a new node (0.0553, 0.0621) in the 2d-tree . . . 358

13.13 Query images and retrieval of images acc ording to various

algorithms with respect to COIL-100 database . . . . . . . . 363

13.14 Query images and retrieval of images acc ording to various

algorithms with respect to Oxford ﬂower-102 database . . . 3 63

13.15 Average F-score values and accuracy values for image cate-

gories in the COIL-100 database . . . . . . . . . . . . . . . . 364

13.16 Average F-score values and accuracy values for image cate-

gories in the Oxford ﬂower-102 database . . . . . . . . . . . 364

14.1 Block diagram of general image storage sy stem . . . . . . . 371

14.2 The basic ﬂow of image compression coding . . . . . . . . . 372

14.3 The stages of the proposed HID system . . . . . . . . . . . . 374

14.4 The block diagram for the encoder of the proposed HID image

compression . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5

14.5 The block diagram for the decoder of the proposed HID image

compression . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5

14.6 2D-DCT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377

14.7 DCT basis functions . . . . . . . . . . . . . . . . . . . . . . 377

14.8 Zig-zag scan . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 78

14.9 Huﬀman encoding . . . . . . . . . . . . . . . . . . . . . . . . 380

14.10 Tested medical images . . . . . . . . . . . . . . . . . . . . . 382

14.11 Quality metrics of the proposed compression technique over

some other ﬁltering methods . . . . . . . . . . . . . . . . . . 386

14.12 Comparison of PSNR over diﬀerent blocks of images at diﬀer-

ent thre shold values . . . . . . . . . . . . . . . . . . . . . . . 386

14.13 Reconstructed images when the threshold TH = 10 . . . . . 387

14.14 Reconstructed images when the threshold TH = 80 . . . . . 387

14.15 Comparison of P SNR value and the bpp of the propose d r e-

constructed image for various image blo cks . . . . . . . . . . 38 9

14.16 Results of tested images for the proposed HID method and

the existing methods SVD, WDR, and JPE G . . . . . . . . 390

14.17 Comparison of PSNR value and the CR of the proposed re-

constructed image over other existing methods . . . . . . . . 391

List of Tables

1.1 Performance analysis for diﬀerent watermarked images using

GA watermarking system. . . . . . . . . . . . . . . . . . . . 21

1.2 Performance analysis for diﬀerent watermarked images using

GA watermarking system. . . . . . . . . . . . . . . . . . . . 27

1.3 Performance analysis for diﬀerent watermarked images using

BFOA watermarking sy stem. . . . . . . . . . . . . . . . . . 27

2.1 Control parameters of CCA. . . . . . . . . . . . . . . . . . . 42

2.2 Best results after applying the CCA using Kapur’s function

to the set of benchmark images. . . . . . . . . . . . . . . . . 46

2.3 Best results of applying the CC A using Otsu’s function to the

set of benchmark images. . . . . . . . . . . . . . . . . . . . . 47

2.4 p-values produced by Wilcoxon’s test c omparing Otsu vs. Ka-

pur over PSNR. . . . . . . . . . . . . . . . . . . . . . . . . . 51

2.5 Compariso n of CCA and MTEMO using K apur’s method. . 52

2.6 Compariso n of CCA with MTEMO using Otsu’s method. . 5 3

2.7 ANOVA test over the CCA and MTEMO based on Kapur’s

metho d. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

2.8 ANOVA test over the CCA and MTEMO based on Otsu’s

metho d. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

3.1 Deviation of lines in diﬀerent angles in both square and hexag-

onal grids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

3.2 Deviation of circles in diﬀerent radii in both square and hexag-

onal grids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

3.3 Deviation of ellipse in both squa re and hexagonal grids. . . 82

4.1 Typical values of tissue’s electrical conductivity. . . . . . . . 96

4.2 EIDORS parameter conﬁgurations. . . . . . . . . . . . . . . 98

4.3 Parameters used for genetic algo rithms. . . . . . . . . . . . . 103

4.4 Parameters used to simulate annealing. . . . . . . . . . . . . 107

4.5 Mutation equations. . . . . . . . . . . . . . . . . . . . . . . 1 15

5.1 No attack vs. imperceptibility and robustness values. . . . . 151

5.2 Spatial attacks and their attack index. . . . . . . . . . . . . 151

5.3 Spatial attacks and their attack index. . . . . . . . . . . . . 153

xvii

xviii List of Tables

6.1 Statistics for the sphere evaluation after 50 runs, using s = 6,

r = 0.95, for diﬀerent rotation angles θ = {π/6, π/4, π/2} and

maximum iteration number k

max

= {50, 100, 400, 700, 1000}. 173

6.2 Statistical analysis of the sphere function evaluation after 50

runs, using s = 6, θ = π/4, for diﬀerent convergence rates r =

{0.20, 0.50, 0.90, 0.95, 0.99} and maximum iteration numbers

max

= {50, 100, 400, 700, 1000 }. . . . . . . . . . . . . . . 174

6.3 Statistics for the f

(2)

sphere

(x) assessment after 50 runs for a

diﬀerent c ombination of k

max

and s, with S(0.95, θ = π/4). 176

6.4 Sphere function evaluation for diﬀerent k

max

and s values

using the φ-model w ith S(0 .95, π/4), for the two random vari-

ables ϕ

= U (0, 1) and ϕ

= N (0.5, 0.25), after 50 runs. . . . 179

6.5 Comparative analy sis of the proposed and the comparative

metho ds, using statistical measures. . . . . . . . . . . . . . . 196

6.6 Comparative analy sis of the proposed and the comparative

metho ds, using statistical measures. . . . . . . . . . . . . . . 197

9.1 Calculated scaling and embedding parameters from test im-

ages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

9.2 Similarity measure between origina l and extracted watermark. 256

9.3 Extracted watermark’s calculated PSNR and NCC values for

gray scale test image House. . . . . . . . . . . . . . . . . . . 258

9.4 Extracted watermark calculated PSNR and NCC after various

attacks from color test image Lena. . . . . . . . . . . . . . . 260

10.1 Comparison in terms of PSNR (DB) resulting from various

ﬁlters at diﬀerent noise intensities for Moon image. . . . . . 2 87

10.2 Comparison in terms of SSIM resulting from various ﬁlters at

diﬀerent noise intensities for Moon image. . . . . . . . . . . 288

10.3 Comparison in terms of IEF resulting from various ﬁlters at

diﬀerent noise intensities for Moon image. . . . . . . . . . . 289

10.4 Comparison in terms o f IQI resulting from various ﬁlters at

diﬀerent noise intensities for Moon image. . . . . . . . . . . 290

10.5 Comparison among various ﬁlters in terms of SSIM at diﬀerent

noise intensities for Lena image. . . . . . . . . . . . . . . . . 290

10.6 Comparison among various ﬁlters in terms of SSIM at diﬀerent

noise intensities for Moon image. . . . . . . . . . . . . . . . 291

10.7 Comparison among various ﬁlters in terms of SSIM at diﬀerent

noise intensities for Peppers image. . . . . . . . . . . . . . . 29 1

11.1 Total number of optimum features selected using FGFS. . . 314

11.2 Precisio n rate of prop osed system. . . . . . . . . . . . . . . . 318

11.3 Recall rate of proposed system. . . . . . . . . . . . . . . . . 319

剩余436页未读，继续阅读

atarashi

粉丝: 0

生物启发计算在图像与视频处理中的创新理论与应用

《高性能CUDA C编程: 创建和调试性能CUDA C》

致敬经典：Redmond-Inspired WordPress主题与Windows XP精神

禅意高清壁纸新体验：Serenity Zen-inspired Wallpapers HD -crx插件

Bio-Inspired Computing - Theories and Applications_ 9th International Conference

Innovations in Bio-Inspired Computing and Applications

Bio-Inspired Computation and Applications in Image Processing

A bio-inspired structure for vibration isolation

Nature-Inspired Computing and Optimization

樽海鞘群算法Mirjalili发表原论文Salp Swarm Algorithm:A bio-inspired optimizer for engineerin

biomimetics-biologically-inspired computing(生物启发算法)

最新资源