Image segmentation apparatus and method

A method that classifies image data within each of a plurality of windows of the image data, comprising:

• receiving the image data;
• determining a video average within each of a plurality of windows of the image data, wherein each window includes a plurality of subwindows, comprising:
• counting at least one of peaks and valleys within a window of the input image data, wherein the window includes a plurality of subwindows;
• determining an image value for each pixel in each window,
• summing image values within each of the subwindows of each window,
• counting a total number of pixels in each subwindow,
• averaging the image value in each subwindow based on the summation of the image values and the total number of pixels for that subwindow,
• finding, for each window, a maximum average image value of all the subwindows of that window, and
• finding, for each window, a minimum average image value of all the subwindows of that window;
• determining first windows of the image data where the video average of the window is below a first threshold;
• evaluating any such first windows of the image data for peaks;
• determining, for each window, a local roughness of the input image data based on a difference between an absolute value of the maximum average image value and an absolute value of the minimum image average value for that window;
• classifying the input image data of the window based on the count of the one of the peaks and the valleys counted in the counting step and the determined local roughness;
• outputting a signal based on the classification of the input image data.

An apparatus that classifies image data within each of a plurality of windows of the image data, comprising:

• an input device that receives the image data;
• an evaluation device that determines a video average within each window of the image data and outputs at least a signal corresponding to the video average for each window, wherein each window includes a plurality of subwindows, comprising:
• an image value device that determines an image value for each pixel in each window,
• a summation device that determines a summation of image values within each of the subwindows of each window,
• a counting device that counts a total number of pixels in each subwindow,
• an average image value device that determines an average image value in each subwindow based on the summation of the image values and the total number of pixels for that subwindow,
• a maximum average image value device that finds, for each window, a maximum average image value of all the subwindows of that window, and
• a minimum average image value device the finds, for each window, a minimum average image value of all the subwindows of that window;
• and
• a local roughness device that determines, for each window, a local roughness of the input image data based on a difference between an absolute value of the maximum average image value and an absolute value of the minimum image average value for that window;
• a peak/valley counter device that counts one of peaks and valleys within the window of the input image data;
• and
• a processor that receives the signal output from the evaluation device for each window, that sets a first threshold and a second threshold based on the video averages for the plurality of windows, that evaluates windows for valleys where the video average is greater than the first threshold, and that evaluates windows for peaks where the video average is less than the second threshold.