US 7,405,736 B2 · Efficient perceptual/physical color space conversion

Abstract

An imaging or other sensory reproduction system efficiently converts image or other sensory data between a perceptual color space (e.g., the sRGB color space) and a physical color space (unity gamma) or other perceptual/physical sensory models that are related by an expression involving a computationally expensive exponential function. The imaging system calculates exponential functions that can be composed from computationally inexpensive operations, such as square root, square, reciprocal, as well as multiplications and/or additions and subtractions. These computationally less expensive functions are then combined, such as in a weighted and/or offset mean, summation or difference to approximate the computationally expensive exponential function. The imaging system evaluates the expression using the approximation to efficiently yield the converted image data. The efficient conversion between perceptual and physical color spaces allows operations, such as blending and anti-aliasing, to be performed in the physical color space before display of a perceptual color space image.

Claims15

Dependent claims

A computer-readable storage medium encoded with a computer program for effecting efficient conversion on a computer processor having a single instruction, multiple data floating point operation instruction set between sensory data represented in a non-linear scale sensory data model and representation in a linear scale sensory data model, where the non-linear and linear scale sensory data models are related by a conversion function involving a computationally expensive exponential power function, the software program code comprising:

first program code evaluating a first exponential function of plural items of the sensory data using single instruction, multiple data instructions for at least one of square, square root and reciprocal operations;
second program code evaluating a second exponential function of the plural items of the sensory data using single instruction, multiple data instructions for at least one of square, square root and reciprocal operations;
third program code combining the first and second exponential functions of the plural items to approximate the computationally expensive exponential power function of the plural items;
and
fourth program code evaluating the conversion function on the plural items using the approximation of the computationally expensive exponential power function of the plural items produced by the third program code to effect conversion of the plural items between the data models.

2 more elements

A method of effecting efficient conversion on a computer processor having a single instruction, multiple data floating point operation instruction set between sensory data represented in a non-linear scale sensory data model and representation in a linear scale sensory data model, where the non-linear and linear scale sensory data models are related by a conversion function involving a computationally expensive exponential power function, the method comprising:

evaluating a first exponential function of plural items of the sensory data using single instruction, multiple data instructions for at least one of square, square root and reciprocal operations;
evaluating a second exponential function of the plural items of the sensory data using single instruction, multiple data instructions for at least one of square, square root and reciprocal operations;
combining the first and second exponential functions of the plural items to approximate the computationally expensive exponential power function of the plural items;
and
evaluating the conversion function on the plural items using the approximation of the computationally expensive exponential power function of the plural items produced by the combining the first and second exponential functions to effect conversion of the plural items between the data models.

2 more elements

A sensory data processing system for effecting efficient conversion between sensory data represented in a non-linear scale sensory data model and representation in a linear scale sensory data model, where the non-linear and linear scale sensory data models are related by a conversion function involving a computationally expensive exponential power function, the method comprising:

a data storage buffer for storing sensory data;
a processor having a single instruction, multiple data floating point operation instruction set, and programmed to:
evaluate a first exponential function of plural items of the sensory data using single instruction, multiple data instructions for at least one of square, square root and reciprocal operations;
evaluate a second exponential function of the plural items of the sensory data using single instruction, multiple data instructions for at least one of square, square root and reciprocal operations;
combine the first and second exponential functions of the plural items to approximate the computationally expensive exponential power function of the plural items;
and
evaluate the conversion function on the plural items using the approximation of the computationally expensive exponential power function of the plural items produced by the combination of the first and second exponential functions to effect conversion of the plural items between the data models.

4 more elements

Dependencies are inferred from the claim text.

Efficient perceptual/physical color space conversion

Life of this patent

Efficient perceptual/physical color space conversion

Life of this patent

Abstract

Claims15