An accurate visibility metric is produced with relatively few calculations.
This invention relates to the devices and methods for the measurement and/or for the specification of the perceptual intensity of a visual image, or the perceptual distance between a pair of images. Grayscale test and reference images are processed to produce test and reference luminance images. A luminance filter function is convolved with the reference luminance image to produce a local mean luminance reference image. Test and reference contrast images are produced from the local mean luminance reference image and the test and reference luminance images, respectively, and are followed by the application of a contrast sensitivity filter. The resulting images are combined according to mathematical prescriptions to produce a Just Noticeable Difference, JND value, indicative of a Spatial Standard Observer (SSO). Advantages of an SSO include a simple and efficient design that produces an accurate visibility metric with relatively few calculations.
Some embodiments of the present invention use a Minkowski sum directly over filtered image pixels. This technique avoids the need for complicated spatial frequency filter banks, with a corresponding gain in simplicity and computational efficiency. A particular form of Contrast Sensitivity Filter (CSF) is used, which combines radial- and oblique-effect filters. This permits accurate visibility predictions of the visibility of the oblique patterns, such as half-toning and rasterizing artifacts. Viewing distance and image resolution are jointly treated in an advantageous manner in some embodiments of this invention. The use of this feature causes the computed value of image visibility to be substantially independent of image resolution (except to the extent that the resolution actually alters the visibility of the information in the image). A window function is advantageously employed in such a manner as to represent the reduction in visibility with distance from the observer’s region of fixation. Additionally, the use of convolving operations, along with the window function, is also useful to make it feasible to simulate the scanning of an image by the eye of the observer. Pooling the data accumulates the visibility over the scan.
When images are located near a border region, it may occur that the border has a markedly different intensity (typically darker) than that of the image and the general image background. In such cases, it is advantageous to introduce special procedures for handling border effects. Two examples are presented. One includes at least a portion of the border into the definition of “image” leading to an enhanced image that is then processed by the SSO. Another approach is to attenuate the image contrast near the border.
This work was done by Andrew B. Watson of Ames Research Center.