| 摘要 |
PURPOSE:To emphasize the density of an image with no conspicuous noise when the image is processed despite the overlap caused between even density parts of images, by calculating the density variance corresponding to the image density and performing weighting to decide a window curve. CONSTITUTION:A dispersion value is obtained by a 1st arithmetic means 4 from the density value of the pixels divided into a matrix form. This obtained dispersion value is added to dispersion data stored in a memory 6 which stores dispersed histogram data corresponding to the density of each pixel via the means 4. Based on the result of this addition, the integrated value of probability which is the sum of these results is calculated by a 2nd arithmetic means 5. Then a window curve is decided, based on said integrated value of probability and the conversion of density is carried out by a 3rd arithmetic means 8. Thus it is possible to decide a weighted window curve in response to evenness of density and to properly emphasize the density of the image. |
