Browse Prior Art Database

Adaptive Contrast Ranging for Images

IP.com Disclosure Number: IPCOM000083974D
Original Publication Date: 1975-Aug-01
Included in the Prior Art Database: 2005-Mar-01
Document File: 4 page(s) / 74K

Publishing Venue

IBM

Related People

Wong, KY: AUTHOR

Abstract

The problem of contrast ranging arises when an image to be printed has a greater contrast dynamic range than the printer is capable of producing.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 4

Adaptive Contrast Ranging for Images

The problem of contrast ranging arises when an image to be printed has a greater contrast dynamic range than the printer is capable of producing.

In Fig. 1, the horizontal axis represents the intensity scale of the input image, while the vertical axis represents the intensity of the output. The problem of contrast ranging amounts to selecting suitable conversion curves for picture inputs of different contrast ranges.

For images whose picture information is in the range of R(1), curve C(1) is a reasonable choice. Likewise if the input is limited to an intensity range of R(2), curve C(2) seems to be reasonable also. However, if the input has a wide dynamic intensity range, such as R(3), the straight line curve C(3) may not be a good choice. This is due to the logarithmic effect of the human vision, i.e., the human eye senses a brightness level change when k = Delta I/I exceeds certain threshold value, where Delta I stands for change of intensity. The threshold value of k has been determined to be 0.14.

If the slope of the straight line C(3) is less than unity (which is the case when the printing ink is not as dark as the darkest area of the input), changes in intensity level that would be large enough in the original would not be large enough in the output picture, for detectable brightness changes by human vision. Thus, there is a loss of picture information content from the input to the output picture. However, some amount of loss of picture fidelity is unavoidable, since the printing device does not have as much contrast range as the original.

In Fig. 2 a family of curves are drawn with different slopes Alpha and different intercepts Beta with the horizontal line passing through y(min). The slope Alpha and the intercept Beta can be varied dynamically as the gray-scale picture data is being supplied for printing.

Human vision is more sensitive to a certain range of spatial frequencies, which has been determined to be in the range of 0.6-12 cycles per degree (or equivalently 2.4-48 cycles/in at 14" of viewing distance). This suggests that Alpha and Beta in Fig. 2 should be tuned to correspond to the picture information in the passband of the human vision.

In Fig. 3, the spatial frequency picture information may be presented serially as time domain information. This data is passed through a filter 10 corresponding to the human eye response model. The filter output v(f) then goes through a threshold detector 11, whereby it is given an 'on' signal the absolute value of v(f) > v(T) ( v(T) is the threshold voltage).

The signal from the threshold detector controls the timing and the duration of an averaging circuit, which estimates the DC component of the scanner...