Apparatus and method for lossy compression using dynamic domain quantization

An apparatus for reducing the dynamic range of input numbers representing a physical quantity, the apparatus comprising:

• a) an input register for successively receiving and storing said input numbers;
• b) a plurality of mapping registers for storing mapping weights of a mapping function;
• c) a processor in communication with said input memory and said plurality of mapping registers;
• d) a processor-readable medium on which is stored instructions operable to direct said processor to:
• i) determine the arithmetic differences between the contents of said input register and said mapping registers;
• ii) determine a nearest mapping register associated with the least arithmetic difference;
• iii) increase or decrease the contents of said nearest mapping register in proportion to said least arithmetic difference;
• iv) replace the contents of the input register with successive sample numbers and repeat the above steps for a plurality of sample numbers representative of said physical quantity;
• v) perform said mapping function on each of said input numbers of said sample set to produce corresponding output numbers.

An apparatus for compressing an initial set of numbers representing a graphic image, said initial set of numbers having an initial dynamic range, the apparatus including:

• a) an input register for receiving and storing a sample number from a sample set of sample numbers representing said graphic image;
• b) a plurality of registers for storing mapping weights of a mapping function;
• c) a processor in communication with said input memory and said plurality of mapping registers;
• d) a processor-readable medium on which is stored instructions operable to direct said processor to:
• i) produce a reduced set of numbers representing said graphical image, by:
• 1) determining the arithmetic differences between the contents of said input register and said mapping registers;
• 2) determining a nearest mapping register is associated with the least arithmetic difference;
• 3) increasing or decreasing the contents of said nearest mapping register in proportion to said least arithmetic difference;
• 4) replacing the contents of the input register with successive sample numbers and repeating the above steps for a plurality of samples representative of said physical quantity;
• 5) successively performing said mapping function on each of said input numbers to produce a sequence of corresponding output numbers, said corresponding output numbers acting as said reduced set of numbers;
• and
• ii) to produce an edge representative series of numbers representing changes and locations of changes in said reduced set of numbers by:
• 1) determining the arithmetic differences between successive output numbers and, when a difference is non-zero, producing a number pair including said output number and a number representing the location of said output number in said output number sequence.