公开(公告)号：US20130266237A1

公开(公告)日：2013-10-10

申请号：US13500070

申请日：2010-12-16

申请人： Bo Li ,  Hai-Heng Cao ,  Yi-hong Wen ,  Hong-xu Jiang

发明人： Bo Li ,  Hai-Heng Cao ,  Yi-hong Wen ,  Hong-xu Jiang

IPC分类号： H04N7/30

CPC分类号： H04N19/00818 ,  H04N19/124 ,  H04N19/154 ,  H04N19/18 ,  H04N19/63

摘要： A wavelet coefficient quantization method using a human visual model in an image compression process is provided, which is particularly suitable for remote sensing image compression. A wavelet-domain visual quantization model is obtained through experiments based on human visual characteristics, so as to reflect the relation between distortion of wavelet coefficient blocks and human visual characteristics. The model includes a luminance component, a masking component and a frequency component, where the luminance component is calculated by low frequency coefficients after the kth level wavelet transform, the masking component is calculated by high frequency coefficients of the second and third levels, and the frequency component is calculated by a statistical method. The method may be used in combination with any mainstream wavelet compression method such as EZW, SPIHT or EBCOT.

摘要翻译： 提供了一种在图像压缩处理中使用人类视觉模型的小波系数量化方法，特别适用于遥感图像压缩。 通过基于人类视觉特征的实验获得小波域视觉量化模型，以反映小波系数块的失真与人类视觉特征之间的关系。 该模型包括亮度分量，掩蔽分量和频率分量，其中在第k级小波变换之后通过低频系数计算亮度分量，通过第二和第三水平的高频系数来计算掩蔽分量， 频率分量通过统计学方法计算。 该方法可以与任何主流小波压缩方法（如EZW，SPIHT或EBCOT）结合使用。

公开(公告)号：US08655092B2

公开(公告)日：2014-02-18

申请号：US13500070

申请日：2010-12-16

申请人： Bo Li ,  Hai-Heng Cao ,  Yi-hong Wen ,  Hong-xu Jiang

发明人： Bo Li ,  Hai-Heng Cao ,  Yi-hong Wen ,  Hong-xu Jiang

IPC分类号： G06K9/00

CPC分类号： H04N19/00818 ,  H04N19/124 ,  H04N19/154 ,  H04N19/18 ,  H04N19/63

摘要： A wavelet coefficient quantization method using a human visual model in an image compression process is provided, which is particularly suitable for remote sensing image compression. A wavelet-domain visual quantization model is obtained through experiments based on human visual characteristics, so as to reflect the relation between distortion of wavelet coefficient blocks and human visual characteristics. The model includes a luminance component, a masking component and a frequency component, where the luminance component is calculated by low frequency coefficients after the kth level wavelet transform, the masking component is calculated by high frequency coefficients of the second and third levels, and the frequency component is calculated by a statistical method. The method may be used in combination with any mainstream wavelet compression method such as EZW, SPIHT or EBCOT.

摘要翻译： 提供了一种在图像压缩处理中使用人类视觉模型的小波系数量化方法，特别适用于遥感图像压缩。 通过基于人类视觉特征的实验获得小波域视觉量化模型，以反映小波系数块的失真与人类视觉特征之间的关系。 该模型包括亮度分量，掩蔽分量和频率分量，其中在第k级小波变换之后通过低频系数计算亮度分量，通过第二和第三水平的高频系数来计算掩蔽分量， 频率分量通过统计学方法计算。 该方法可以与任何主流小波压缩方法（如EZW，SPIHT或EBCOT）结合使用。