A
Adaboost, see Adaptive Boosting (Adaboost)
Adaptive Boosting (Adaboost), 34, 36, 44; see also Classifiers
Adaptive data-hiding method, 220; see also Image steganographic schemes, VQ-based
Adobe PhotoshopTM, 297
Affine-invariant interest point detector, 52
Alattar’s method, 120
ANN, see Artificial neural network (ANN)
AR coefficients, see Autoregressive coefficients (AR coefficients)
Arnold’s cat map, 100; see also Pseudo-random pixel rearrangement algorithm
image rearrangement, 111
multiplication time complexity, 110
Artificial neural network (ANN), 301
Autoregressive coefficients (AR coefficients), 301
B
Bag-of-feature method, 55, 56; see also Image categorization
Bayesian classifier, 58
Bag of words (BoW), 7
model, 32
Bagging, 33
Bayesian classifier, 58
Bayesian formula, 58
BBCM, see Block-based chaotic map (BBCM)
Behavior modeling of human objects, see Human behavior modeling
BER, see Bit error rate (BER)
Bi-normal separation (BNS), 32
Bit error rate (BER), 248, 260
Bits per pixel (bpp), 116, 136
Black NC (BNC), 150
Block-based chaotic map (BBCM), 86, 87; see also Digital watermarking
block size and local image similarity, 88
for changing pixel, 86
transformed image and significant coefficients, 86, 87
Block-based fragile watermarking system, 181; see also Digital watermarking; Fragile watermarking
cyclic redundancy checksum, 183
extension to color images, 189–191
lookup table generation, 184–185
methodology comparison, 183–184
proposed method, 185
related work, 183
tampered blocks recovery, 189
vector quantization attack, 183
Block size and local image similarity, 88
BNC, see Black NC (BNC)
BNS, see Bi-normal separation (BNS)
Boolean functions, 149
Boosting, 33; see also Adaptive Boosting (Adaboost)
BoW, see Bag of words (BoW)
bpp, see Bits per pixel (bpp)
Broadcast media
CF removal, 7
duplicate videos sharing, 4
C
CA, see Correspondence analysis (CA)
Camera-captured image, 301
CDF, see Cumulative distribution function (CDF)
CF chronological distribution, 9
Asahi Breweries, 11
Daihatsu Tanto Exe, 12
CFs, see Commercial films (CFs)
CFS, see Correlation-based feature selection (CFS)
Chan et al.’s scheme, 128–130, 131
Chang et al.’s codebook clustering-based scheme, 132
for embedding one bit of secret data, 132
for embedding two bits of secret data, 133
limitation, 134
Chaos, 86, 314; see also Chaos-based hash function
merits of, 314
to protect image security, 343; see also Self-synchronizing method
theory, 86
Chaos-based hash function, 314
changeable-parameter and self-synchronizing keystream, 317
collision resistance analysis, 321–322
detection hash performance, 318
diffusion and confusion statistic analysis, 319–321
grouping and processing, 316
hash value sensitivity, 318–319
implementation and flexibility, 324–325
localization hash performance, 322–323
modification detection and localization capabilities, 317
padding, division, and preprocessing, 315–316
parallel mode, 317
preliminaries, 314
speed analysis, 324
Chaotic cryptography, 318
Chaotic image encryption scheme, 343–344; see also Self-synchronizing method
Chaotic map, 87
Chaotic systems, 314
toral automorphism, 86
CHI, see Chi-square measure (CHI)
Chi-square measure (CHI), 32, 35
Chromosome, 170, 259; see also Genetic algorithm (GA); Genetic algorithm-based breaking methodology
length and QF values, 171
pixel value adjustment, 260
Classifiers, 33, 34; see also Video concept classification
Adaboost, 36
in bag-of-feature method, 58
Bayesian, 58
DT, 36
JRip, 36
k-NN, 37
learning, 36
MCA-based, 37
Naive Bayes, 36
SVM, 58
traditional, 33
clustering, 135
clustering-based scheme, 132
construction algorithm, 55, 56
DES-encrypted, 239
encrypted, 238
small state, 232
state, 134
Commercial film mining, 7; see also Commercial films (CFs); Multimedia duplicate mining
application, 7
Asahi Breweries CF chronological distribution, 11
Daihatsu Tanto Exe CF chronological distribution, 12
knowledge discovery, 9, 10, 11
recurring fragment pairs, 9
temporal recurrence hashing algorithm, 8–9, 10
Commercial films (CFs), 4, 7–8
cluster, 8
fragment, 8
mining, see Commercial film mining monitoring, 4, 7
recurring fragment pairs, 9
removal, 7
sequence, 8
Compression ratio, 136
Computational forensics, 327, see Video forensics
Computer images, 203
Content-based image retrieval, 52
Content-based video copy detection, 26
Copies, see Near duplicates copy–cover, 302
copy–cover image forgery detection, 298, 302
DCT domain method, 305
experimental results, 305
lexicographical sorting, 304
PCA domain method, 304
spatial domain method, 305
statistical domain method, 305
Copy-paste (CP), 166
Correlation-based feature selection (CFS), 34, 35
Correspondence analysis (CA), 37
Cover-object, 258
CP, see Copy-paste (CP)
CRC, see Cyclic redundancy checksum (CRC) Criminisi’s algorithm, 276, 277; see also Image inpainting
confidence term, 278
data term, 278
drawback of, 284
priority value, 278
single iteration of, 277
Cross dimensional indexing, 18, 19–21
Cryptographic hash function, 314; see also Chaos-based hash function
Cryptography, 219, 245, 257; see also Hashing; Image steganography; Steganography
Cumulative distribution function (CDF), 40
Cyclic redundancy checksum (CRC), 169–170, 183
D
Data encryption, 216
Data hiding, 115; see also Reversible data-hiding techniques
capacity, 116
conflicting factors in, 117
difference image-based, 124
nibble-based, 121
Data mining, 31
WEKA, 36
Data transmissions security, 216
db, see Decibels (db)
DBN, see Dynamic Bayesian network (DBN)
DCT, see Discrete Cosine Transform (DCT)
DE, see Differential evolution (DE)
DE-based techniques, see Difference expansion-based techniques (DE-based techniques)
Decibels (db), 174
Decision Tree (DT), 33
classification results of, 45
comparison of best F1_scores, 47
Degree of freedom, 40
Description length (DL), 36
DFT, see Discrete Fourier transform (DFT)
Difference expansion-based techniques (DE-based techniques), 117; see also Reversible data-hiding techniques
experimental results, 121
nibble-based data-hiding scheme, 121
overflow and underflow problem, 120
Thodi and Rodriguez’s PEE scheme, 118–120
Difference image-based data-hiding scheme, 121; see also Data hiding; Reversible data-hiding techniques
Difference-of-Gaussian detection method (DoG detection method), 52
Differential evolution (DE), 246, 248, 255; see also Differential evolution-based algorithm
application procedure, 251
flowchart, 250
method of, 248
vector, 252
Differential evolution-based algorithm, 245; see also Differential evolution (DE)
and image steganography, 248
objective function, 248
stego-image generation, 247
VSS methodology breaking for, 247–248
Digital images; see also Digital watermarking; Image categorization; JPEG images authentication; Reversible data-hiding techniques
data hiding in images, 216
manipulation, 297
security of, 343; see also Self-synchronizing method
software used for, 297
Digital images authentication, see JPEG images authentication
Digital media types, 215
Digital signature, 165
Digital steganographic schemes, 215; see also Steganography; Vector quantization (VQ)
based on VQ codebook modification, 221
data-embedding procedure, 230
experimental results, 223–227, 230–235, 238
SOC algorithm, 229
visually imperceptible image scheme based on VQ, 235–243
by VQ and LSB modification, 221–227
in VQ-compressed domain of image, 227–235
Digital steganography, 216, 245; see also Digital steganographic schemes
requirements, 68
types, 68
Digital watermarking, 67, 85, 141, 166, 202, 219, 257, 299; see also Dual watermarking scheme; Image steganography; JPEG images authentication; Visible watermarking; Watermark
block-based chaotic map, 86–88
complete fragile, 142
domain-based watermarking techniques, 85, 86
experimental results, 92
fragile watermarking, 166
LBP operator and its extensions, 143, 161–162
LSB method, 142
multilevel watermarking based on LBP operators, 156
patchwork methods, 142
progress, 142
proposed watermarking algorithm, 89
robustness experiment, 94, 95, 96
spatial, 148
spatial-domain techniques, 166
spread-spectrum modulation techniques, 142
threshold effects under JPEG compression, 97
Discrete Cosine Transform (DCT), 85; see also Integer wavelet transform (IWT); Digital images; Discrete Fourier transform (DFT); Discrete wavelet transform (DWT)
based methods (DCT-based methods), 68
extracting strategy, 93
Discrete Cosine Transform domain method, 305
Discrete Fourier transform (DFT), 142; see also Discrete Cosine Transform (DCT)
Discrete wavelet transform (DWT), 142; see also Discrete Cosine Transform (DCT)
DL, see Description length (DL)
DoG detection method, see Difference-of-Gaussian detection method (DoG detection method)
Domain-based watermarking techniques, 85; see also Digital watermarking
frequency-domain watermarking, 86
wavelet-domain watermarking algorithm, 86
DT, see Decision Tree (DT)
Dual-watermarking scheme, 182; see also Fragile watermarking
mining, 5
DWT, see Discrete wavelet transform (DWT)
Dynamic Bayesian network (DBN), 374
E
EA, see Evolutionary algorithm (EA)
EBR, see Error bit rate (EBR)
Electronic watermark, see Digital watermark EM approach, see Estimation maximization approach (EM approach)
Error bit rate (EBR), 150, 151
and contrast adjustment, 155, 159
and JPEG compression, 155, 156, 159
and luminance linear transformation, 154, 159
and rotation, 156
Error measures, 251
ES, see Evolution strategy (ES)
Estimation maximization approach (EM approach), 368
Evolution strategy (ES), 246
Evolutionary algorithm (EA), 248; see also Genetic algorithm (GA)
Exact copy, see Exact duplicate Exact duplicate, 5
Exact duplicate mining, 5; see also Multimedia duplicate mining (MDM); Near-duplicate mining
Exclusive OR operator (XOR operator), 149, 197, 248, 263, 316
location map, 128
EzStego, 210; see also Steganography software
F
Fallahpour and Sedaaghi’s method, 124; see also Histogram shifting-based techniques (HS-based techniques)
False acceptance ratio (FAR), 172, 174
under attacks, 178
FAR, see False acceptance ratio (FAR)
Fast Fourier transform (FFT), 207; see also Multilayer fractional Fourier transform (MLFFT)
Fourier-moment-based image registration, 75
in MATLAB, 207
Fast quantization
FDSS, see Frequency-domain steganalytic systems (FDSS)
Feature extraction, 52; see also Image categorization
affine-invariant interest point detector, 52
group activity recognition, 375, 377
Harris-Laplace detector, 52
high-level, 34
for image and video processing, 359
in multimedia content analysis, 358, 375
rotation matrix, 53
social network analysis, 376
TRECVID videos, 38
Feature selection techniques, 32, 34; see also Video concept classification
categories, 34
CFS, 35
embedded methods, 34
I-Relief method, 35
multivariate methods, 35
ranker method, 36
univariate methods, 35
wrappers, 34
Feedback sets, 15
FFT, see Fast Fourier transform (FFT)
Fingerprint, 5; see also Near-duplicate mining
in criminal investigations, 328–329
errors, 5
temporal information capturing, 8
Fitness function, 170, 259, 260; see also Genetic algorithm (GA)
Fixed embedding methods, 221
Fixed-length chromosome procedure, 171–172
Foreground, 70
Forensic science training process, 328–329; see also Computational forensics; Video forensics
Fourier-based image registration, 72; see also Iterative refinement process; Visible watermarking
algorithm, 75
Fourier-moment-based image registration algorithm, 75
log-polar transform, 72
moment matching approach, 74–75
rigid transform relationship, 72–74
translation vector, 73
Fragile watermarking, 166, 182; see also Block-based fragile watermarking system; Digital watermarking
blockwise-dependent techniques, 182
blockwise-independent techniques, 182
complete, 142
drawback, 300
dual-watermarking scheme, 182–183
Lin et al scheme, 182
uncompressed image authentication, 166
Frequency domain, 204; see also Spatial domain; Image steganography
reversing, 130
Frequency-domain steganalytic systems (FDSS), 209; see also Spatial-domain steganalytic system (SDSS)
GA-based breaking algorithm on, 264
Frequency-domain watermarking, 86
G
GA, see Genetic algorithm (GA)
Gaussian blurring, 309
Gaussian integer, 100, 106; see also Pseudorandom pixel rearrangement algorithm
algorithm for finding Gaussian integer generators, 103
cryptoimmunity of the rearrangement algorithm, 108
generator, 103
left-to-right binary exponentiation algorithm, 102
modulo (mod) operation, 101
multiplication algorithm, 101
natural number, 100
proof of algorithm validity, 106
Gaussian process dynamical model (GPDM), 359
group activity recognition, 377–378
trajectory learning, 365
Genetic algorithm (GA), 207, 208, 246, 259; see also Genetic algorithm-based breaking methodology
complexity of, 270
on the JFDSS, 266
operations, 261
recombination strategy of, 261–262
stego-image generated by, 267
Genetic algorithm-based breaking algorithm on SDSS, 262; see also Genetic algorithm-based breaking methodology
experimental results, 268
stego-image generation, 262
Genetic algorithm-based breaking methodology, 259; see also Genetic algorithm-based breaking algorithm on SDSS
chromosome containing 64 genes, 259–260
crossover, 261
fitness function, 260
mutation, 261
recombination strategy, 261
reproduction, 260
on VSS, 266
Genetic algorithm-based watermark-embedding method, 167; see also JPEG images authentication
authentication information generation, 168–169, 169
block adjustment, 171
chromosome evaluation objective function, 170–171
chromosome substitution in original block, 170
cyclic redundancy checksum, 169–170
fixed-length chromosome, 171–172
GA adjustment, 170
procedure to adjust block, 169
variable-length chromosome, 172, 173
watermark-embedding procedure, 167–168
GIMPTM(GNU Image Manipulation Program), 297
GNU Image Manipulation Program, see GIMPTM(GNU Image Manipulation Program)
GPDM, see Gaussian process dynamical model (GPDM)
H
Haar digital wavelet transform (HDWT), 125; see also Reversible data-hiding techniques
Chan et al.’s scheme, 128–130, 131
Harris-Laplace detector, 52
Hash functions, see Fingerprint Hashing, 314; see also Chaos-based hash function
conventional algorithm, 325
temporal recurrence algorithm, 8–9
HCPs, see High-curvature points (HCPs) HDWT, see Haar digital wavelet transform (HDWT)
HGPDM, see Hierarchical Gaussian process dynamical model (HGPDM)
Hidden Markov model-based method (HMM-based method), 380
Hiding capacity, 116, 239; see also Data hiding
Alattar’s method, 120
HS-based reversible data-hiding method, 124
Kamstra and Heijmans’ scheme, 126
solving problem of inefficient, 236
Hierarchical Gaussian process dynamical model (HGPDM), 361, 364, 371; see also Gaussian process dynamical model (GPDM); Human motion change detection
classification algorithm, 368
creation, 362
experimental results, 369–371, 372
focus, 362
HGPLVM, 361
human interaction modeling, 361
human motion training in, 364
with particle filter, 366
with particle filter in testing phase, 366
particle filtering, 362–364, 381
Hierarchical Gaussian process latent variable model (HGPLVM), 359
High-curvature points (HCPs), 333
edge features and gradient information histogram, 334
High-density sampling, 5
Histogram of orientfed gradient (HOG), 358, 360
Histogram shifting-based techniques (HS-based techniques), 122; see also Reversible data-hiding techniques
difference image-based data-hiding scheme, 124–125
experimental results, 125
Fallahpour and Sedaaghi’s method, 124
HMM-based method, see Hidden Markov
model-based method (HMM-based method)
HOG, see Histogram of oriented gradient (HOG)
Horizontal wavelet transform (HWT), 118
of cover image, 126
HS-based techniques, see Histogram shifting-based techniques (HS-based techniques)
Hu’s scheme, see Virtual image cryptosystem
Human activity recognition methods, 373; see also Human group behavior analysis
intergroup and intragroup interaction, 374
non-target-oriented methods, 373–374
target-oriented methods, 374
tracking, 374
Human behavior modeling, 358; see also Human group behavior analysis; Human motion change detection
challenges, 359
discussions and future directions, 381–382
feature categorization, 359
Human group behavior analysis, 372; see also Human activity recognition methods; Human motion change detection; Human behavior modeling
adaptive mean-shift tracking, 375–376
approaches for, 374
BEHVAE dataset results, 379–380
centrality histogram, 377
conditional probability calculation, 378
experimental results, 378
GPDM creation, 378
GPDM selection with highest conditional probability, 378
group activity recognition framework, 375
human group behavior modeling, 377
IDIAP dataset results, 380–381
middle-level features, 361
model parameters initialization, 378
motion histogram, 376
small group activity recognition, 377–378
small group clustering, 376
social network-based feature extraction, 376–377
statistical shape theory, 374
Human motion change detection, 361; see also Hierarchical Gaussian process dynamical model; Human group behavior analysis; Human behavior modeling
human motion trajectory data extraction, 365
importance weight equation, 369
latent space motion classification and change detection, 363, 367–368
leaf node latent space particle, 364
model parameters initialization, 362
particle filter framework initialization, 363, 366–367
particle filter tracking and update, 368
probabilistic mapping determination, 364
and tracking framework, 363
trained running latent space, 367
trajectory learning model on leaf node, 365
weight updating, 364
Human visual system (HVS), 68, 142
HVS, see Human visual system (HVS)
HWT, see Horizontal wavelet transform (HWT)
I
IBPC method, see Intersection-based pixels collection method (IBPC method)
IDCT, see Inverse Discrete Cosine Transform (IDCT)
IEM, see Information entropy maximization (IEM)
IG, see Information gain (IG); InGroup (IG)
Image
distance measure, 6
domain, see Spatial domain
editing software, 297
Image authentication
fragile watermarking scheme for, 182
watermarking for, 299
for uncompressed images, 166
Image categorization, 51; see also Bag-of-feature method
applications, 51
classification rate for each run, 61
confusion matrix with SVM, 62
feature extraction and description, 52–54
K-means clustering and nearest neighbor quantization, 56
linear soft-margin SVMs, 58
major runs, 60
nearest-neighbor classification, 58
result comparison, 61
Image compression, 203; see also Discrete Cosine Transform (DCT)
application of, 216
codebook, 228
encoding and decoding modules, 217
JPEG, 207
lossless compression, 203
PSNR, 253
Image forgery, 298
detection, 307
Gaussian blurring, 309
image splicing, 300
Image inpainting, 275, 277; see also Criminisi’s algorithm
algorithm and associated problems, 279–281
artifact reduction, 285
conventional, 337
dynamic searching range, 283–285
experimental results, 285
image filling processing, 280, 283
of large missing area, 275–276
man-made images evaluation, 294–295
natural image results, 288–293
patch-based, 338
proposed algorithm, 281
searching range expansion process, 284
structure priority value, 281–283
target region shape influence analysis, 286–288
Image manipulation, 206
Image-processing operations, 151, 152, 158, 162
Image quality measure (IQM), 246, 258
stego-image on, 263
Image splicing, 300
camera-captured images and step-like edge transitions, 301
trace detection algorithm, 301–302
Image steganographic schemes, VQ-based, 216, 219; see also Digital steganographic schemes
adaptive data-hiding method, 220
on compressed domain of images, 220
data embedding by LSB substitution, 218–219
fixed embedding methods, 221
search-order-coding algorithm, 217–218
Image steganography, 216, 221; see also Digital steganographic schemes; Digital watermarking; Steganalysis
classifications, 204
comparison, 252
on compressed domain of images, 220
DE algorithm and, 248
purpose of, 219
scheme, 220
Index table, 133
compact, 220
underlined index, 132
Inductive learning algorithms, 36
Information entropy maximization (IEM), 37
Information gain (IG), 32
Information hiding, 67, 215; see also Digital watermarking; Steganography
cover image pixel value, 219
formula, 237
issue in modified LSB method, 68
InGroup (IG), 379
overlapped central histogram of, 377
Integer wavelet-based multiple logo-watermarking schemes, 300
Integer wavelet transform (IWT), 117; see also Discrete Cosine Transform (DCT)
reversible data-hiding method, 130
Intersection-based pixels collection method (IBPC method), 88
Inverse Discrete Cosine Transform (IDCT), 208
Invertible data-hiding scheme, see Reversible data-hiding techniques
Invisible watermarking, 68; see also Digital watermark
IQM, see Image quality measure (IQM)
I-Relief method, see Iterative relief method (I-Relief method)
Isophote, 280; see also Criminisi’s algorithm
Iterative refinement process, 75; see also Fourier-based image registration; Visible watermarking
binary edge image data, 75
Fourier-based image registration algorithm with, 77
Iterative relief method (I-Relief method), 35
IWT, see Integer wavelet transform (IWT)
J
Java Distributed Evolutionary Algorithms Library (JDEAL), 251
JDEAL, see Java Distributed Evolutionary Algorithms Library (JDEAL)
Jeffrey-Lindley paradox, 41
Joint Photographic Experts Group (JPEG), 246, 298
JPEG, see Joint Photographic Experts Group (JPEG)
JPEG images authentication, 165; see also Genetic algorithm-based watermarkembedding method
authentication algorithm, 172–174
digital signature, 165
digital watermarking, see Digital watermarking
GA-based watermark-embedding method JPEG compression, 166
JPEG procedure, 167
JPeg-JSteg software (JSteg software), 211–212; see also Steganography software
JSteg software, see JPeg-JSteg software (JSteg software)
K
Kamstra and Heijmans’ scheme, 126; see also Reversible data-hiding techniques
cover image HWT, 126
embedding result of, 128
local variance, 127
maximum hiding capacity achieved, 126
neighboring pixel pair selection, 127
pixel pairs expansion results, 128
purpose of, 126
window in L image, 127
k-means algorithm, 55
k-Nearest Neighbor (k-NN), 33
algorithm, 58
k-NN, see k-Nearest Neighbor (k-NN)
L
Laplacian-of-Gaussian (LoG), 52
magnitude of, 53
LBG, see Linde-Buzo-Gray (LBG)
LBP operators, see Local binary pattern operators (LBP operators)
LDA, see Linear discriminative analysis (LDA)
Least significant bits (LSB), 68
chi-square attack, 258
modification technique, 235
steganalysis, 209
steganographic techniques, 206
S-Tools, 210
substitution, 112, 204, 205, 218–219; see also Digital watermarking
Left-to-right binary exponentiation
algorithm, 102; see also Gaussian integer
License plate recognition methods, 330; see also Video forensics
Linde-Buzo-Gray (LBG), 217
codewords generation, 222, 228
Linear discriminative analysis (LDA), 376
Lin et al scheme, 182
Local binary pattern operators (LBP operators), 142, 143, 147, 161–162
Boolean functions, 149
completed modeling of, 148
EBR and contrast adjustment, 155
EBR and JPEG compression, 155, 156
EBR and luminance linear transformation, 154
EBR and noise, 154
EBR and rotation, 156
experimental results and analysis, 150
extended circle neighborhoods, 145
and its extensions, 143
image-processing operations, 152
improved, 147
locally rotation invariant pattern, 145
local region, 148
luminance, 154
multilevel watermarking based, see Multilevel watermarking
original LBP operator extensions, 143, 144–148
pattern rotation, 146
patterns, 146
PSNR and intensity factor, 153
sign function, 144
spatial watermarking based on, 148
symmetric texture patterns, 146
tamper detection and location, 153
texture spectrum histogram, 144
uniform, 145
watermark embedding, 149–150, 155
watermark extraction algorithm, 150
LoG, see Laplacian-of-Gaussian (LoG)
Lookup table generation, 184–185, 185
Lossless compression techniques, 238
Lossless data-hiding scheme, see Reversible data-hiding techniques
Lossy compression images, 246, 258
Low image distortion, 116
LSB, see Least significant bits (LSB)
M
Market research on media, 7
MBR, see Minimum boundary rectangle (MBR)
MCA, see Multiple correspondence analysis (MCA)
MDL, see Minimum description length (MDL)
MDM, see Multimedia duplicate mining (MDM)
Mean gray-level embedding method (MGLE method), 221
MER, see Message error rate (MER)
Message error rate (MER), 93
MGLE method, see Mean gray-level embedding method (MGLE method)
Minimum boundary rectangle (MBR), 283
Minimum description length (MDL), 37
Minimum span tree (MST), 376
MLFFT, see Multilayer fractional Fourier transform (MLFFT)
Model-based motion detection methods, 359
Modulo (mod) operation, 101
Moment matching approach, 74–75
Most significant bit (MSB), 170, 182–183
Motion activity, 334; see also Video forensics
Motion descriptors, 333; see also Video forensics
object-based, 334
Motion detection algorithms, 359
Motion flow-based video retrieval scheme, 335
Motion recognition
change detection framework, 371
optical flow, 360
small human groups, 359
MSB, see Most significant bit (MSB)
MST, see Minimum span tree (MST)
Multilayer fractional Fourier transform (MLFFT), 72
Multilevel watermarking, 156; see also Local binary pattern operators (LBP operators)
double-level watermarking, 156–160
EBR and contrast adjustment, 159
EBR and JPEG compression, 159
EBR and luminance linear transformation, 159
EBR and noise, 158
examples, 161
extracted watermarks, 157, 158
four-level watermarking, 160
LBP pattern, 157
Multimedia content analysis, 358, 375; see also Human behavior modeling
feature selection, 359
Multimedia data mining, 4
Multimedia duplicate mining (MDM), 4
advantage, 4
application, 4
duplicate mining methods selection criterion, 5
exact duplicate mining, 5
filler material detection, 4
knowledge-discovery applications, 4
main elements, 26
near-duplicate mining, 6
news story retrieval and threading, 12–17
scaling up video mining, indexing for, 17
Multimedia semantic analysis
approach to overcome challenges in, 33
distinctive feature set, 359
traditional classifiers, 33
Multiple correspondence analysis (MCA), 33, 37
classification results, 47
correlation, 40
dual-model classification, 37, 42–43
feature-value pairs, 38
indicator matrix, 38
principal inertia, 39
reliability, 40
symmetric map, 39
training data set transactions, 42
Multiplication algorithm, 101
N
NA, see Not available (NA)
Natural number, 100; see also Gaussian integer
NB, see Naive Bayes (NB)
NC, see Normalized correlation (NC)
Near-duplicate mining, 6; see also Exact duplicate mining; Multimedia duplicate mining (MDM)
Near duplicates, 6
News story, 13; see also News story retrieval tracking, 4
News story retrieval, 12, 13; see also Multimedia duplicate mining
background, 12
eigenvector centrality, 15
experimentation evaluation, 16–17
feedback sets, 15
limitation, 12
must-link constraint, 15
news stories example, 14
news story reranking process, 13
PageRank, 16
queries selected for experimentation, 16
visual near duplicates, 14
News topic, 13; see also News story retrieval
Nibble-based data-hiding scheme, 121; see also Difference expansion-based techniques (DE-based techniques)
Ni et al’s HS scheme, 122; see also Histogram shifting-based techniques (HS-based techniques)
cover image pixel values, 122
Pixel values after HS, 124
stego image pixel values, 124
Non-Blum Gaussian primes, 101; see also Gaussian integer
Normalized correlation (NC), 150
Not available (NA), 240
NPCR, see Number of Pixel Change Rate (NPCR)
Number of Pixel Change Rate (NPCR), 353
O
Object categorization, 52, 54; see also Image categorization
cost function, 54
histogram, 55
k-means clustering, 54
vector quantization method, 55
Object recognition, 52; see also Image categorization
local descriptor-based algorithm, 333, 334
target objects detection, 335
OIV, see Original index value (OIV)
Optical flow, 360
Optimization method, 248; see also Genetic algorithm (GA)
Original index value (OIV), 218
P
Paint, NetTM, 297
Pair-wise nearest-neighbor embedding method (PNNE method), 221
Partial differential equation-based methods (PDE-based methods), 337
Patch-based methods, 337
Patchwork methods, 142
PCA, see Principal component analysis (PCA)
PDE-based methods, see Partial differential equation-based methods (PDE-based methods)
Peak signal-to-noise ratio (PSNR), 116, 150, 253
and image distortion, 174, 251
and intensity factor, 153
watermarked images, 160, 174, 191
PEE scheme, see Prediction error expansion scheme (PEE scheme)
Percentage of recovery (PR), 191
Piecewise Linear Chaotic Map (PWLCM), 314
Pixel, 203
Pixel rearrangement algorithm, 100, 103
based on Gaussian integers, 103–104
extraction of watermark applied with, 105
watermarking with, 105
PNNE method, see Pair-wise nearest-neighbor embedding method (PNNE method)
Pol, see Point of Interest (Pol)
Point of Interest (Pol), 19, 20
Population, 170, 259; see also Genetic algorithm (GA)
Positive training data instances, 32
Posture sequence retrieval technique, 338
PPCA, see Probabilistic principal component analysis (PPCA)
PR, see Percentage of recovery (PR)
Prediction error expansion scheme (PEE scheme), 118–120; see also Difference expansion-based techniques (DE-based techniques)
Primary colors, 203
Principal component analysis (PCA), 32; see also DCT
domain method, 303, 304–305, 310
as feature selection method, 33
negative rates, 308
Principal inertia, 39
Probabilistic principal component analysis (PPCA), 363
Pseudo-random pixel rearrangement algorithm, 100; see also Watermark
advantages, 112
algorithm validity, proof of, 106–108
comparison to Arnold’s cat map chaos transformation, 110–111
example in image watermarking, 111–112
pixel rearrangement algorithm, 103
rearrangement algorithm cryptoimmunity, 108–110
PSNR, see Peak signal-to-noise ratio (PSNR)
PWLCM, see Piecewise Linear Chaotic Map (PWLCM)
PyML, 56
Q
QBE, see Query-By-Example (QBE)
QF, see Quality factor (QF)
Quality factor (QF), 88
Query-By-Example (QBE), 12
R
Radial basis function (RBF), 365
RBF, see Radial basis function (RBF)
Red, green, and blue (RGB), 203
in digital cameras, 301
histogram, 366
spectrum, 349
REF methods, see Relief filter methods (REF methods)
Reference register (RR), 86; see also Digital watermarking
approaches of collecting pixels, 88
IBPC, 88
subimages similarity by IBPC, 89
Region of interest (ROI), 137, 182
Relief filter methods (REF methods), 34
Repeated incremental pruning to produce error reduction (RIPPER), 36
Reversible data-hiding techniques, 115, 137; see also Data hiding
applications, 116
based on IWT, 130
based on joint neighboring coding, 136
based on SMVQ, 134
basic requirements, 116
Chan et al’s HDWT-based scheme, 128–130
Chang et al’s codebook clustering-based scheme, 132–134
compression domain, 131, 134–136
experimental results, 131, 136
fragile technique, 116, 117, 137
hiding capacity, 116
high-capacity, 135
Kamstra and Heijmans’ scheme, 126–128
new integer-to-IWT-based, 131
publications, 116
research directions, 137
semifragile technique, 116, 117
spatial domain, 117
threshold-embedding technique, 130
transformed domain, 117, 125, 130
visual quality, 116
in VQ compression domain, 135
for VQ indices, 135
Reversible watermarking algorithm, 300
RGB, see Red, green, and blue (RGB)
RIPPER, see Repeated incremental pruning to produce error reduction (RIPPER)
algorithm, 300
ROI, see Region of interest (ROI)
Rotation, scaling, and translation-invariant (RST-invariant), 329; see also Object recognition
RR, see Reference register (RR)
RST-invariant, see Rotation, scaling, and translation-invariant (RST-invariant)
S
Sampling rate, 5
Scaled conjugate gradient (SCG), 362
Scale-invariant feature (SIFT), 19, 32, 53, 332
comparison of recall, 25
features, 360
3D-SIFT, 361
Scaling up video mining, indexing for, 17, 26; see also Multimedia duplicate mining
advantage, 17
background, 17
cross-dimensional index structure, 18
cross-dimensional indexing, 19–21
drawbacks, 17
dual space indexing, 19
indexing process, 22
temporal consistency, 23
SCG, see Scaled conjugate gradient (SCG)
SD, see Standard deviation (SD)
SDSS, see Spatial-domain steganalytic system (SDSS)
Search-order coding algorithm (SOC algorithm), 217–218, 229, 234–235
Search point (SP), 218
Security of key, 323–324; see also Chaos-based hash function
Self-synchronizing method, 343
chaotic systems, 344
chosen-plaintext/known-plaintext attack, 348–350
correlation coefficient, 352
encryption result, 351
original image partition, 345
original scheme, 344
performance analysis, 350
stream cipher, 345
Semifragile watermark, 167
drawback, 300
Semi-fragile watermarking, 166, 182
Sequential forward selection (SFS), 32
SFS, see Sequential forward selection (SFS)
Side-match vector quantization (SMVQ), 134
encoding process, 231
traditional, 232
SIFT, see Scale-invariant feature (SIFT)
Signature-based motion detection methods, 359
Significant coefficients, 86–87
Singular value decomposition (SVD), 39
Small group activity analysis, 359, 379
SMVQ, see Side-match vector quantization (SMVQ)
SNA, see Social network analysis (SNA)
SOC algorithm, see Search-order coding algorithm (SOC algorithm)
Social network analysis (SNA), 376
SP, see Search point (SP)
Space-time interest points (STIP), 360
Spatial domain, 204; see also Frequency domain; Steganography
method, 305
negative rate, 309
Spatial-domain steganalytic system (SDSS), 209
algorithm for stego-image on, 264
GA-based breaking algorithm on, 262
procedure for embedded-image on, 264
Spread-spectrum modulation techniques, 142
SP value, see Structure priority value (SP value)
SSD, see Sum of squared differences (SSD)
Standard correspondence analysis, 37
Standard deviation (SD), 94
Statistical analysis, 208–209; see also Steganalysis
Statistical domain method, 305, 308, 309
Statistical significance, see p-value
Steganalysis, 203, 208, 246, 258; see also Steganography
categories, 209
software, 213
statistical analysis, 209
steganographer, 203
visual analysis, 209
Steganographer, 203
Steganography, 99, 201, 202, 206, 245–246; see also Digital watermarking; Steganalysis
algorithms and transformations, 207–208, 258
classifications, 204
common approaches, 204
digital, 246
and digital watermarking, 202
frequency domain, 204
fundamental requirement, 246
genetic algorithm, 208
in historic times, 246
image processing, 204
image representation, 203
intent of, 201
linguistic, 246
LSB substitution, 204
software, 209
spatial domain, 204
steganalytic systems, 258
stego-object, 258
Steganography software, 209
EzStego, 210
JPeg-JSteg, 211
steganalysis software, 213
S-Tools, 210
StegDetect, 213; see also Steganography software
Stego image, 115
Stego-image generation, 262; see also Genetic algorithm-based breaking algorithm on SDSS
on JFDSS, 265
Stego-key, 206
Stego-object
generation, 247
STIP, see Space-time interest points (STIP)
S-Tools, 210; see also Steganography software
Structure priority value (SP value), 281; see also Image inpainting
filling order and confidence term, 282
ratio of edge points, 282
Sum of squared differences (SSD), 278
Support vector machine (SVM), 55, 58, 360
one-against-all, 59
SVD, see Singular value decomposition (SVD)
SVM, see Support vector machine (SVM)
T
Tamper detection, 153, 186–189
Tampered blocks recovery, 189
Target/positive concept, 32
Temporal recurrence hashing, 10
Texture synpaper method, 275
Thodi and Rodriguez’s PEE scheme, 118–120
Tian’s DE scheme, 118, 120; see also Difference expansion-based techniques (DE-based techniques)
Toral automorphism, 86
Transform domain, see Frequency domain
Translation vector, 73
Transparent watermarking, see Invisible watermarking
TV commercials, see Commercial films (CFs)
U
UACI, see Unified Average Changing Intensity (UACI)
Unexpandable pixel, 120
Unified Average Changing Intensity (UACI), 353, 354
V
Variable-length chromosome procedure, 172, 173
Variable-length coding techniques (VLC techniques), 217–218
Variant environmental conditions, 359
Variant steganalytic systems, 246
Vector quantization (VQ), 182, 216–217; see also Digital steganographic schemes; Side-match vector quantization (SMVQ)
attack, 183
based image steganographic schemes, 216–217
based steganographic technique, 227
codebook modification and image steganographic scheme, 221
coding system, 217
compression system components, 216
encoding and decoding modules, 217
image compressed domain and image
steganographic scheme, 227–235
image compression, 216
method, 55
parts, 217
PyML, 56
reversible data-hiding method, 132
SOC algorithm, 229
visually imperceptible image steganographic scheme, 235–243
Video, 6
features for analysis, 360
Video concept classification, 31, 48; see also Multiple correspondence analysis (MCA)
Adaboost classification results, 44
best F1_scores comparison, 47
concepts for evaluation, 44
data imbalance, 32
discriminative learning framework, 37, 38
DT classification results, 45
experiments and results, 43
JRip classification results, 45
k-NN classification results, 46
NB classification results, 46
Video forensics, 328; see also Video inpainting
candidate character positions and estimation, 331
critical component, 328
event detection process, 335–337
existing technologies, 330
license plate image recognition, 330–332
mining for criminal evidence, 329
motion flow-based video retrieval system, 328
people counting in videos, 338–339
required techniques, 338
RST-invariant object recognition, 332–333
source character images into character template images, 331
trajectory of moving objects, 328, 333–337
video objects recognization, 339–340
Virtual image cryptosystem, 219–220, 236
Visible document watermarking, 70
Visible image watermark, 69
Visible watermark removal, 77
for pictures, 78
for text images, 79
watermark detection, 79
watermarked area and neighboring region intensity relationship, 78
Visible watermarking, 68; see also Digital watermarking
background, 70
critical point, 68
experiments and discussions, 79–82
foreground, 70
Fourier-based image registration, 72–75
iterative refinement process, 75–77
limitation, 68
reversible, 68
visible document watermarking, 70
visible image watermark, 69
watermark removal, 70, 71, 77–79
watermarked documents, 71
Visual analysis, 209; see also Steganalysis
Visual features, 358; see also Semantic gap
Visually imperceptible image steganography, 235; see also Digital steganographic schemes
based on VQ, 236
compressed information volume, 237
image-embedding procedures at transmitter, 237
image extraction at receiver, 237
secret images’ extraction, 238
transmitter input and output, 236
Visual steganalytic system (VSS), 246, 262
DE-based methodology for breaking, 247
GA-based breaking algorithm on, 266
Stego-image generation on, 263
VLC techniques, see Variable-length coding techniques (VLC techniques)
VQ, see Vector quantization (VQ)
VQ-based data-hiding technique, 231
VSS, see Visual steganalytic system (VSS)
W
WalkingTogether (WT), 379
Watermark, 99, 141, 299, 245; see also Digital watermarking; Pseudo-random pixel rearrangement algorithm; Visible watermarking
detection, 79
in frequency domain, 142
general procedure of, 299
for image authentication, 299
integer exponentiation modulo prime, 100
security of, 300
techniques, 300
Watermark embedding, 69–70, 149, 168, 185–186
GA-based watermark-embedding, 167
in spatial domain, 143
Watermarked area and neighboring region intensity relationship, 78
Wavelet, 208
Wavelet-domain watermarking algorithm, 86
WEKA, 36
White normalized correlation (WNC), 150
WINDOW operations, 162
WNC, see White normalized correlation (WNC)
WT, see WalkingTogether (WT)
X
XOR operator, see Exclusive OR operator (XOR operator)
Z