The analysis of methods for image contrast enhancement, which use global transformations, 
is carried out. These methods are Look-up-Table, histogram transformation, 
histogram warping and image contrast improvement based on differential gray-levels histogram. 
The drawbacks of last one are shown. A new class of methods for image contrast enhancement 
based on the characteristics of local regions (relative local contrast, standard deviation, 
trimmed average, median, min, max and order statistic) is proposed. The values of these 
local region characteristics are the base for construction of histogram and corresponding 
to it global transformation function. Some examples, which illustrate the effectiveness 
of proposed methods, are shown. Proposed approaches can be used not only for equalization, 
but also for histogram hyperbolization, exponential and logarithmic transformation. 
Another way for extension of this class of methods is using of transformations in moving window.