TY - JOUR
T1 - Assessment method of fusion image quality in wavelet domain structural similarity
AU - Zhang, Yong
AU - Jin, Weiqi
PY - 2012/6
Y1 - 2012/6
N2 - For the fusion image quality evaluation, the relationship between fusion image quality assessment and image quality assessment is analyzed, and the general expression of fusion image quality assessment is provided. Meanwhile, it is explained that the key of fusion image quality assessment is to construct the reference image that is inexistent actually. Based on the above analysis and the space domain structural similarity evaluation method, wavelet decomposes input source images and fusion image respectively and structure reference image wavelet coefficient is obtained based on the wavelet decomposition coefficient of input source images. Afterwards, by weighing different levels of wavelet frequency band according to the human eye sensation characteristics, the wavelet domain structural similarity metrics to the whole fusion image are created. Comparing the mutual information and space domain structural similarity evaluation method with the subjective assessment index using target detectable ability, detail resolvable ability and whole comfortable ability as a reference, experimental results show that wavelet domain structural similarity method agrees with human visual features and its objective results. It is more consistent with the subjective assessment results than traditional methods.
AB - For the fusion image quality evaluation, the relationship between fusion image quality assessment and image quality assessment is analyzed, and the general expression of fusion image quality assessment is provided. Meanwhile, it is explained that the key of fusion image quality assessment is to construct the reference image that is inexistent actually. Based on the above analysis and the space domain structural similarity evaluation method, wavelet decomposes input source images and fusion image respectively and structure reference image wavelet coefficient is obtained based on the wavelet decomposition coefficient of input source images. Afterwards, by weighing different levels of wavelet frequency band according to the human eye sensation characteristics, the wavelet domain structural similarity metrics to the whole fusion image are created. Comparing the mutual information and space domain structural similarity evaluation method with the subjective assessment index using target detectable ability, detail resolvable ability and whole comfortable ability as a reference, experimental results show that wavelet domain structural similarity method agrees with human visual features and its objective results. It is more consistent with the subjective assessment results than traditional methods.
KW - Assessment
KW - Fusion
KW - Human visual system
KW - Image processing
KW - Wavelet domain structural similarity
UR - http://www.scopus.com/inward/record.url?scp=84866249559&partnerID=8YFLogxK
U2 - 10.3788/CJL201239.s109007
DO - 10.3788/CJL201239.s109007
M3 - Article
AN - SCOPUS:84866249559
SN - 0258-7025
VL - 39
JO - Zhongguo Jiguang/Chinese Journal of Lasers
JF - Zhongguo Jiguang/Chinese Journal of Lasers
IS - SUPPL.1
M1 - s109007
ER -