Data hiding for vector quantization images using mixed-base notation and dissimilar patterns without loss of fidelity

Chang, Chin‐Chen; Lin, Chih-Yang; Hsieh, Yi-Pei

doi:10.1016/j.ins.2011.12.025

Cited by 22 publications

(15 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The comparative results show that our proposed method is better than Chang's new method [22] in the embedding capacity and bit rate. The proposed method embeds more secret data in each index averagely with the lower bit rate compared with Chang's new method [22] and the image quality of the proposed method is the same as Chang's method [22]. Table 13 shows the comparative results with other SMVQ-based data hiding methods.…”

Section: The Comparison With the Jnc-based Data Hiding Methodsmentioning

confidence: 90%

“…The bit rate is almost the same as Yang's method [18] and the image quality is absolutely the same as Yang's, but our embedding capacity is almost three times as much as the capacity in Yang's method [18]. Table 12 gives the comparative results with Chang's new data hiding method [22]. The results of proposed method shown in Table 12 are the results of under hiding.…”

Section: The Comparison With the Jnc-based Data Hiding Methodsmentioning

confidence: 92%

“…Yang et al [21] proposed reversible data hiding based on side match and hit pattern for VQ-compressed images. Chang et al [22] proposed a reversible data hiding scheme based on mixed base notation and dissimilar pattern strategies for vector quantization. Chang put the dissimilar codewords into same cluster and embedded secret data based on the mixed base notation according to the clusters.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A new lossless data hiding method based on joint neighboring coding

Pan¹,

Hu²,

Ma³

et al. 2015

Journal of Visual Communication and Image Representation

View full text Add to dashboard Cite

Section: The Comparison With the Jnc-based Data Hiding Methodsmentioning

confidence: 90%

Section: The Comparison With the Jnc-based Data Hiding Methodsmentioning

confidence: 92%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A new lossless data hiding method based on joint neighboring coding

Pan¹,

Hu²,

Ma³

et al. 2015

Journal of Visual Communication and Image Representation

View full text Add to dashboard Cite

“…Ressalte-se que, além da compressão de sinais, há um amplo espectro de aplicações para QV, como, por exemplo, esteganografia [7,8], marca d'água digital [9,18], identificação vocal [25] e classificação de sinais de voz com patologias [27].…”

Section: Figura 1: Exemplo De Qv De Imagemunclassified

Uma Alternativa de Aceleração do Algoritmo Fuzzy K-Means Aplicado à Quantização Vetorial

Madeiro¹,

Galvão²,

Ferreira³

et al. 2012

TCAM

View full text Add to dashboard Cite

Resumo. Compressão de sinais, marca d´água digital e reconhecimento de padrões são exemplos de aplicações de quantização vetorial (QV). Um problema relevante em QV é o projeto de dicionários. Neste trabalho, é apresentada uma alternativa de aceleração do algoritmo fuzzy K-Means aplicado ao projeto de dicionários. Resultados de simulações envolvendo QV de imagens e de sinais com distribuição de Gauss-Markov mostram que o método proposto leva a um aumento da velocidade de convergência (redução do número de iterações) do algoritmo fuzzy K-Means sem comprometimento da qualidade dos dicionários projetados.Palavras-chave. Quantização vetorial, fuzzy K-Means, codificação de imagens. IntroduçãoA quantização é uma técnica relevante em sistemas de compressão de sinais [12,21]. Seu alvo é a digitalização (discretização) dos valores de amostras de um sinal. Há duas classes gerais de quantização: escalar e vetorial. A primeira consiste em um mapeamento Q : R → C, em que C é um subconjunto do espaço Euclidiano unidimensional R. O número de elementos de C é denominado número de níveis do quantizador, denotado por L. Ao final do processo de quantização, cada amostra poderá ser codificada, por exemplo, em uma palavra-binária de l = log 2 L bits. A quantização vetorial (QV), por sua vez, consiste em um mapeamento Q : R N → W , em que W = { w i ; i = 1, 2, . . . , K} ⊂ R N é denominado dicionário, N é a dimensão do quantizador vetorial e K é o tamanho do dicionário (número de vetores-código). A taxa de codificação da quantização vetorial é dada por 1 N log 2 K, expressa em bits por amostra em codificação de forma de onda de voz e em bits por pixel (bpp) em codificação de imagem. 1 madeiro@poli.br 2

show abstract

“…Literatures [22,11,31] embed data by index mapping techniques based on refer counts of the indices. The idea behind literatures [2,5,9,23,30] is to map the indices based on side match prediction.…”

Section: Introductionmentioning

confidence: 99%

A low distortion reversible data hiding scheme for search order coding of VQ indices

Rahmani

Dastghaibyfard

2014

Multimed Tools Appl

View full text Add to dashboard Cite

In recent years many reversible data hiding schemes for vector quantization (VQ)-compressed images have been developed. However most of them generate non-legitimate codes as output and make the data hiding behavior detectable. Moreover existing schemes with legitimate outputs usually need side information to achieve reversibility. In this paper a reversible data hiding scheme for VQ-compressed images based on search order coding (SOC) is proposed which generate legitimate SOC codes as output. For each index of the input VQ index table which finds a match in its corresponding SOC path, the indices located in the SOC path are clustered, and then SOC codes of some blocks are changed to SOC code of one of the indices in the same cluster to hide data. Some SOC codes are re-encoded using original index value (OIV) codes for recovery purpose. By the strategy of re-encoding, the proposed scheme doesn't need any side information. The proposed scheme adaptively adjusts the criteria of exchangeability of the indices based on the smoothness level of each image block to prevent appearing spots in the stego-image. Experimental results show that the proposed scheme on average could embed about 0.5 bits per index by only about 0.67 dB degradation of image quality for a set of test images includes smooth an complicated images. Also superiority of the proposed scheme when compared with the previous schemes is confirmed from various aspects; hiding capacity, stego-image quality, the average of the bit rate of the output code, and the execution time.

show abstract

Data hiding for vector quantization images using mixed-base notation and dissimilar patterns without loss of fidelity

Cited by 22 publications

References 18 publications

A new lossless data hiding method based on joint neighboring coding

A new lossless data hiding method based on joint neighboring coding

Uma Alternativa de Aceleração do Algoritmo Fuzzy K-Means Aplicado à Quantização Vetorial

A low distortion reversible data hiding scheme for search order coding of VQ indices

Contact Info

Product

Resources

About