Median determination apparatus and method

A method of determining a median of an image data set of size M, the method comprising the steps of:

• receiving each value of the image data set as a weighted set of binary digits;
• adding together a first constant and binary digits of greatest weight of each value of the image data set;
• dividing the result of the adding step by a second constant to generate a binary digit of greatest weight of a median of the image data set;
• calculating for each value a remaining value clamp function for a next lower weight of binary digits;
• employing results of the calculating step to repeat adding and dividing steps for the next lower weight of binary digits to generate a binary digit of next lower weight of a median of the image data set;
• and
• repeating the calculating and employing steps, if necessary, for each lower weight of binary digits of the values.

An image processing apparatus for processing an image comprising a data set of size M, said apparatus comprising:

• means for receiving each value of the data set as a weighted set of binary digits;
• means for adding together a first constant and binary digits of greatest weight of each value of the data set;
• means for dividing the result provided by said adding means by a second constant to generate a binary digit of greatest weight of a median of the data set;
• means for calculating for each value a remaining value clamp function for a next lower weight of binary digits;
• means for employing results provided by said calculating means to repeat adding and dividing for the next lower weight of binary digits to generate a binary digit of next lower weight of a median of the data set;
• and
• means for repeating calculating and employing, if necessary, for each lower weight of binary digits of the values.

An apparatus for computing a median of an image data set of size M, said apparatus comprising:

• means for receiving each value of the image data set as a weighted set of binary digits;
• means for adding together a first constant and binary digits of greatest weight of each value of the image data set;
• means for dividing the result provided by said adding means by a second constant to generate a binary digit of greatest weight of a median of the image data set;
• means for calculating for each value a remaining value clamp function for a next lower weight of binary digits;
• means for employing results provided by said calculating means to repeat adding and dividing for the next lower weight of binary digits to generate a binary digit of next lower weight of a median of the image data set;
• and
• means for repeating calculating and employing, if necessary, for each lower weight of binary digits of the values.