Image processing method and apparatus with pixel conversion using reaction-diffusion

An image processing method comprising:

• carrying out, by using a computer, scale conversion for a first pixel value of each of a plurality of pixels included in an image to generate a second pixel value of each of the plurality of pixels;
• applying, by using the computer, a reaction-diffusion equation including a diffusion element and a reaction element that is set according to at least the number of types of regions to be extracted, to the second pixel value of each of plural pixels within a certain region of the image a predetermined number of times to generate a third pixel value of each of the plurality of pixels included in the image;
• and
• carrying out, by using the computer, scale inverse-conversion that is inverse-conversion of the scale conversion, for the third pixel value of each of the plurality of pixels included in the image to generate a fourth pixel value of each of the plurality of pixels,
• wherein the scale conversion is a processing to linearly convert a pixel value within a value range of the pixel value into a value within a range between −1 and +1, and
• the scale inverse-conversion is a processing to linearly convert the value within the range between −1 and +1 into the pixel value within the value range of the pixel value, and
• a value obtained by carrying out the scale conversion for a threshold for the region extraction within the value range of the pixel value is set as a root of a plurality of roots of the reaction element, which corresponds to a threshold for region extraction.

An image processing apparatus comprising:

• a memory storing data of an image;
• a processing unit configured to execute a process, comprising:
• carrying out scale conversion for a first pixel value of each of a plurality of pixels included in the image to generate a second pixel value of each of the plurality of pixels;
• applying a reaction-diffusion equation including a diffusion element and a reaction element that is set according to at least the number of types of regions to be extracted, to the second pixel value of each of plural pixels within a certain region of the image a predetermined number of times to generate a third pixel value of each of the plurality of pixels included in the image;
• and
• carrying out scale inverse-conversion that is inverse-conversion of the scale conversion, for the third pixel value of each of the plurality of pixels included in the image to generate a fourth pixel value of each of the plurality of pixels,
• wherein a predetermined constant value is set as a root of a plurality of roots of the reaction element, which corresponds to a threshold for region extraction, and
• each of first sections obtained by dividing a value range of the pixel value by a lower limit value and an upper limit value of the pixel value and a threshold for the region extraction within the value range of the pixel value are associated with one of second sections obtained by dividing a range between −1 and +1 by −1, +1 and the predetermined constant value,
• the scale conversion is a first processing to carry out intra-section linear mapping between one first section of the first sections and a second section associated with the one first section, and
• the scale inverse-conversion is a second processing to carry out an inverse processing of the first processing.

An image processing apparatus comprising:

• a memory storing data of an image;
• a processing unit configured to execute a process, comprising:
• carrying out scale conversion for a first pixel value of each of a plurality of pixels included in the image to generate a second pixel value of each of the plurality of pixels;
• applying a reaction-diffusion equation including a diffusion element and a reaction element that is set according to at least the number of types of regions to be extracted, to the second pixel value of each of plural pixels within a certain region of the image a predetermined number of times to generate a third pixel value of each of the plurality of pixels included in the image;
• and
• carrying out scale inverse-conversion that is inverse-conversion of the scale conversion, for the third pixel value of each of the plurality of pixels included in the image to generate a fourth pixel value of each of the plurality of pixels,
• wherein the scale conversion is a processing to linearly convert a pixel value within a value range of the pixel value into a value within a range between −1 and +1, and
• the scale inverse-conversion is a processing to linearly convert the value within the range between −1 and +1 into the pixel value within the value range of the pixel value, and
• a value obtained by carrying out the scale conversion for a threshold for the region extraction within the value range of the pixel value is set as a root of a plurality of roots of the reaction element, which corresponds to a threshold for region extraction.

A computer-readable, non-transitory storage medium storing a program for causing a computer to execute a process, the process comprising:

• carrying out scale conversion for a first pixel value of each of a plurality of pixels included in an image to generate a second pixel value of each of the plurality of pixels;
• applying a reaction-diffusion equation including a diffusion element and a reaction element that is set according to at least the number of types of regions to be extracted, to the second pixel value of each of plural pixels within a certain region of the image a predetermined number of times to generate a third pixel value of each of the plurality of pixels included in the image;
• and
• carrying out scale inverse-conversion that is inverse-conversion of the scale conversion, for the third pixel value of each of the plurality of pixels included in the image to generate a fourth pixel value of each of the plurality of pixels,
• wherein a predetermined constant value is set as a root of a plurality of roots of the reaction element, which corresponds to a threshold for region extraction, and
• each of first sections obtained by dividing a value range of the pixel value by a lower limit value and an upper limit value of the pixel value and a threshold for the region extraction within the value range of the pixel value are associated with one of second sections obtained by dividing a range between −1 and +1 by −1, +1 and the predetermined constant value,
• the scale conversion is a first processing to carry out intra-section linear mapping between one first section of the first sections and a second section associated with the one first section, and
• the scale inverse-conversion is a second processing to carry out an inverse processing of the first processing.

A computer-readable, non-transitory storage medium storing a program for causing a computer to execute a process, the process comprising:

• carrying out scale conversion for a first pixel value of each of a plurality of pixels included in an image to generate a second pixel value of each of the plurality of pixels;
• applying a reaction-diffusion equation including a diffusion element and a reaction element that is set according to at least the number of types of regions to be extracted, to the second pixel value of each of plural pixels within a certain region of the image a predetermined number of times to generate a third pixel value of each of the plurality of pixels included in the image;
• and
• carrying out scale inverse-conversion that is inverse-conversion of the scale conversion, for the third pixel value of each of the plurality of pixels included in the image to generate a fourth pixel value of each of the plurality of pixels,
• wherein the scale conversion is a processing to linearly convert a pixel value within a value range of the pixel value into a value within a range between −1 and +1, and
• the scale inverse-conversion is a processing to linearly convert the value within the range between −1 and +1 into the pixel value within the value range of the pixel value, and
• a value obtained by carrying out the scale conversion for a threshold for the region extraction within the value range of the pixel value is set as a root of a plurality of roots of the reaction element, which corresponds to a threshold for region extraction.