Evaluation of Image Quality Features Via Monotonic Analysis

Abstract : This work introduces novel monotonic analysis to determine whether or not proposed image quality (IQ) measures are consistent with human measured perceptual quality scores. Specifically, the analysis performs a generalized likelihood ratio test over the H1 hypothesis that the IQ measures and the corresponding perceptual measurements are related via a monotonic function versus the null hypothesis that the functional relationship is arbitrary. This paper evaluates six proposed IQ measures against mean opinion scores using the new monotonic analysis. The next generation of night vision goggles and night scopes will fuse image intensified (I2) and long wave infrared (LWIR) to create a hybrid image that will enable soldiers to better interpret their surroundings during nighttime missions. The key to such systems is the determination of the best image fusion algorithm for a specific task. A number of image fusion algorithms have been proposed in the literature. Currently, a scientific evaluation of such algorithms requires extensive and expensive human perception studies to determine how well soldiers can perform a specific task. What is needed is an image quality (IQ) measure than can automatically quantify the utility of image fusion algorithms.

[1]  Richard K. Moore,et al.  Monotonic correlation analysis of image quality measures for image fusion , 2008, SPIE Defense + Commercial Sensing.

[2]  Pramod K. Varshney,et al.  A human perception inspired quality metric for image fusion based on regional information , 2007, Inf. Fusion.

[3]  Rick S. Blum,et al.  Multi-sensor image fusion and its applications , 2005 .

[4]  P.K. Varshney,et al.  Imaging for concealed weapon detection: a tutorial overview of development in imaging sensors and processing , 2005, IEEE Signal Processing Magazine.

[5]  Colin E. Reese,et al.  Comparison of additive image fusion vs. feature-level image fusion techniques for enhanced night driving , 2003, SPIE Optics + Photonics.

[6]  G. Qu,et al.  Information measure for performance of image fusion , 2002 .

[7]  Lorenzo Bruzzone,et al.  Image fusion techniques for remote sensing applications , 2002, Inf. Fusion.

[8]  Vladimir Petrovic,et al.  Objective image fusion performance measure , 2000 .

[9]  Rick S. Blum,et al.  A categorization of multiscale-decomposition-based image fusion schemes with a performance study for a digital camera application , 1999, Proc. IEEE.

[10]  Panos M. Pardalos,et al.  Algorithms for a Class of Isotonic Regression Problems , 1999, Algorithmica.

[11]  Oliver Rockinger,et al.  Image sequence fusion using a shift-invariant wavelet transform , 1997, Proceedings of International Conference on Image Processing.

[12]  Terrance L. Huntsberger,et al.  Wavelet-based sensor fusion , 1993, Other Conferences.

[13]  Michael J. Best,et al.  Active set algorithms for isotonic regression; A unifying framework , 1990, Math. Program..

[14]  Alexander Toet,et al.  A morphological pyramidal image decomposition , 1989, Pattern Recognit. Lett..

[15]  Edward H. Adelson,et al.  The Laplacian Pyramid as a Compact Image Code , 1983, IEEE Trans. Commun..

[16]  Gordon Pledger,et al.  On Consistency in Monotonic Regression , 1973 .

[17]  Zhou Wang,et al.  Image Quality Assessment: From Error Measurement to Structural Similarity , 2004 .

[18]  Allen M. Waxman,et al.  Color Night Vision: Opponent Processing in the Fusion of Visible and IR Imagery , 1997, Neural Networks.

[19]  Keinosuke Fukunaga,et al.  Introduction to Statistical Pattern Recognition , 1972 .