make_noise_image creates an image with noise to be used in perceptual tests. make_noise_image creates an image consisting of a solid background (the value of which is given by the -b option) with a square region of noise added to the center.
More precisely, make_noise_image creates a subband pyramid of depth specified by levels (option -nl) and initializes it to all zeros. Then, noise of variance noise_variance (option -n) is added in a square in the center of the subband specified by option -s. make_noise_image then performs a 2D separable wavelet reconstruction of the subband pyramid (see imgidwt(1) ). Finally, mean value background (option -b) is added to the entire image, and the result is output to imgfile, a PGM image.
The noise added to the image is zero-mean, uniform, and has variance noise_variance, i.e., the range of the noise is -c to c where c = sqrt(3 * noise_variance).
make_noise_image is intended to be used in the determining of perceptual weightings as outlined in Jayant et al. and in Hontcsh et al. make_noise_image can be used to output images at a number of noise levels, with a number of background intensities, and with noise in different subbands of the pyramidal subband decomposition. An observer can then determine, visually, which level of noise gives just-noticeable distortion (JND) in each subband for each background intensity. After a series of such perceptual experiments, the frequency and brightness sensitivity as a function of the subband number (i.e., frequency) may be determined.
Note: If levels is zero, no subband reconstruction is performed. In this case, noise is added to the baseband subband (i.e., the only band), the mean is added, and the result is output directly.
N. Jayant, J. Johnston, and R. Safranek, "Signal Compression Based on Models of Human Perception," Proceedings of the IEEE, vol. 81, pp. 1385-1422, October 1993.
I. Hontsch, L. J. Karam, and R. J. Safranek, "A Perceptually Tuned Embedded Zerotree Image Coder," Proceedings of the IEEE International Conference on Image Processing, vol. 1, pp. 41-44, October 1997.
M. Antonini, M. Barlaud, P. Mathieu, and I. Daubechies, "Image Coding Using Wavelet Transform," IEEE Transactions on Image Processing, vol. 1, pp. 205-220, April 1992.