Image processing device, image processing method, and image processing program

An image processing device comprising:

• an image obtaining unit configured to obtain a Bayer image in a Bayer array constituted by first rows in each of which red and green pixels are alternately arranged in a horizontal direction and second rows in each of which green and blue pixels are alternately arranged in the horizontal direction, the first rows and the second rows being alternately arranged in a vertical direction;
• a frequency decomposing unit configured to decompose the Bayer image, with respect to each of the vertical direction and the horizontal direction, into a low frequency component representing brightness information and a high frequency component representing color information, thereby to generate a plurality of subimages;
• an image processing unit configured to perform at least one of the following processes for image processing of (i) a characteristic amount computation process of computing the amount of characteristic of the Bayer image based on the subimages obtained through frequency decomposition by the frequency decomposing unit, (ii) a correcting process on one or more of the subimages obtained through the frequency decomposition, and (iii) a set of information compression and information decompression processes on one or more of the subimages obtained through the frequency decomposition, in such a manner that processing details of the at least one of the processes for the high frequency component are different from those for the low frequency component;
• and
• a decomposed image combining unit configured to combine the subimages having been processed by the image processing unit, to regenerate a single Bayer image, wherein the image processing unit comprises:
• a filter obtaining unit configured to obtain first and second coefficient groups including a plurality of filter coefficients by which pixel values of pixels constituting each subimage before processed in the at least one of the processes are multiplied;
• and
• a filtering processing unit configured to filter the low frequency component using the first coefficient group obtained by the filter obtaining unit and filter the high frequency component using the second coefficient group obtained by the filter obtaining unit,
• the first coefficient group is obtained based on a first weight monotonously decreasing depending on a distance from a current pixel, and
• the second coefficient group is obtained based on a second weight monotonously decreasing depending on the distance from the current pixel at a decrease rate smaller than that of the first weight, or having no change depending on the distance from the current pixel.

An image processing device comprising:

• an image obtaining unit configured to obtain a Bayer image in a Bayer array constituted by first rows in each of which red and green pixels are alternately arranged in a horizontal direction and second rows in each of which green and blue pixels are alternately arranged in the horizontal direction, the first rows and the second rows being alternately arranged in a vertical direction;
• a first frequency decomposing unit configured to decompose the Bayer image, with respect to each of the vertical direction and the horizontal direction, into a low frequency component representing brightness information and a high frequency component representing color information, thereby to generate a plurality of subimages;
• a second frequency decomposing unit configured to perform one-level or multiple-level frequency decomposition on each of the subimages generated by the first frequency decomposing unit with respect to each of the vertical direction and the horizontal direction and thereby further decomposing each of the subimages into subimages having different frequency bands;
• an image processing unit configured to perform at least one of the following processes for image processing of (i) a characteristic amount computation process of computing the amount of characteristic of the Bayer image based on the subimages obtained through frequency decomposition by the second frequency decomposing unit, (ii) a correcting process on one or more of the subimages obtained through the frequency decomposition, and (iii) an information compression process on one or more of the subimages obtained through the frequency decomposition, in such a manner that processing details of the at least one of the processes for the high frequency component are different from those for the low frequency component;
• and
• an output unit configured to output at least one of (i) the subimages on which the at least one of the processes for image processing has been performed by the image processing unit and (ii) the amount of characteristic.

An image processing device, comprising:

• an image obtaining unit configured to obtain a Bayer image in a Bayer array constituted by first rows in each of which red and green pixels are alternately arranged in a horizontal direction and second rows in each of which green and blue pixels are alternately arranged in the horizontal direction, the first rows and the second rows being alternately arranged in a vertical direction;
• a frequency decomposing unit configured to decompose the Bayer image, with respect to each of the vertical direction and the horizontal direction, into a low frequency component representing brightness information and a high frequency component representing color information, thereby to generate a plurality of subimages;
• an image processing unit configured to perform at least one of the following processes for image processing of (i) a characteristic amount computation process of computing the amount of characteristic of the Bayer image based on the subimages obtained through frequency decomposition by the frequency decomposing unit, (ii) a correcting process on one or more of the subimages obtained through the frequency decomposition, and (iii) an information compression process on one or more of the subimages obtained through the frequency decomposition, in such a manner that processing details of the at least one of the processes for the high frequency component are different from those for the low frequency component;
• and
• an output unit configured to output at least one of (i) the subimages on which the at least one of the processes for image processing has been performed by the image processing unit and (ii) the amount of characteristic, wherein
• the image processing unit comprises:
• a filter obtaining unit configured to obtain first and second coefficient groups including a plurality of filter coefficients by which pixel values of pixels constituting each subimage before processed in the at least one of the processes are multiplied;
• and
• a filtering processing unit configured to filter the low frequency component using the first coefficient group obtained by the filter obtaining unit and filter the high frequency component using the second coefficient group obtained by the filter obtaining unit,
• the first coefficient group is obtained based on a first weight monotonously decreasing depending on a distance from a current pixel, and
• the second coefficient group is obtained based on a second weight monotonously decreasing depending on the distance from the current pixel at a decrease rate smaller than that of the first weight, or having no change depending on the distance from the current pixel.

An image processing method comprising:

• an image obtaining step of obtaining a Bayer image in a Bayer array constituted by first rows in each of which red and green pixels are alternately arranged in a horizontal direction and second rows in each of which green and blue pixels are alternately arranged in the horizontal direction, the first rows and the second rows being alternately arranged in a vertical direction;
• a first frequency decomposing step of decomposing the Bayer image, with respect to each of the vertical direction and the horizontal direction, into a low frequency component representing brightness information and a high frequency component representing color information, thereby to generate a plurality of subimages;
• a second frequency decomposing step of performing one-level or multiple-level frequency decomposition on each of the subimages generated in the first frequency decomposing step with respect to each of the vertical direction and the horizontal direction and thereby further decomposing each of the subimages into subimages having different frequency bands;
• an image processing step of performing at least one of the following processes for image processing of (i) a characteristic amount computation process of computing the amount of characteristic of the Bayer image based on the subimages obtained through frequency decomposition in the second frequency decomposing step, (ii) a correcting process on one or more of the subimages obtained through the frequency decomposition, and (iii) a set of information compression and information decompression processes on one or more of the subimages obtained through the frequency decomposition, in such a manner that processing details of the at least one of the processes for the high frequency component are different from those for the low frequency component;
• and
• first and second decomposed image combining steps of combining the subimages after the first and second frequency decomposing steps and the image processing step, to regenerate a single Bayer image.

An image processing method comprising:

• an image obtaining step of obtaining a Bayer image in a Bayer array constituted by first rows in each of which red and green pixels are alternately arranged in a horizontal direction and second rows in each of which green and blue pixels are alternately arranged in the horizontal direction, the first rows and the second rows being alternately arranged in a vertical direction;
• a frequency decomposing step of decomposing the Bayer image, with respect to each of the vertical direction and the horizontal direction, into a low frequency component representing brightness information and a high frequency component representing color information, thereby to generate a plurality of subimages;
• an image processing step of performing at least one of the following processes for image processing of (i) a characteristic amount computation process of computing the amount of characteristic of the Bayer image based on the subimages obtained through frequency decomposition in the frequency decomposing step, (ii) a correcting process on one or more of the subimages obtained through the frequency decomposition, and (iii) a set of information compression and information decompression processes on one or more of the subimages obtained through the frequency decomposition, in such a manner that processing details of the at least one of the processes for the high frequency component are different from those for the low frequency component;
• and
• a decomposed image combining step of combining the subimages after the image processing step, to regenerate a single Bayer image, wherein
• the image processing step comprises:
• a filter obtaining step of obtaining first and second coefficient groups including a plurality of filter coefficients by which pixel values of pixels constituting each subimage before processed in the at least one of the processes are multiplied;
• and
• a filtering processing step of filtering the low frequency component using the first coefficient group obtained in the filter obtaining step and filtering the high frequency component using the second coefficient group obtained in the filter obtaining step,
• the first coefficient group is obtained based on a first weight monotonously decreasing depending on a distance from a current pixel, and
• the second coefficient group is obtained based on a second weight monotonously decreasing depending on the distance from the current pixel at a decrease rate smaller than that of the first weight, or having no change depending on the distance from the current pixel.

A non-transitory computer-readable recording medium recording thereon a computer program causing a computer to execute the following steps of:

• an image obtaining step of obtaining a Bayer image in a Bayer array constituted by first rows in each of which red and green pixels are alternately arranged in a horizontal direction and second rows in each of which green and blue pixels are alternately arranged in the horizontal direction, the first rows and the second rows being alternately arranged in a vertical direction;
• a first frequency decomposing step of decomposing the Bayer image, with respect to each of the vertical direction and the horizontal direction, into a low frequency component representing brightness information and a high frequency component representing color information, thereby to generate a plurality of subimages;
• a second frequency decomposing step of performing one-level or multiple-level frequency decomposition on each of the subimages generated in the first frequency decomposing step with respect to each of the vertical direction and the horizontal direction and thereby further decomposing each of the subimages into subimages having different frequency bands;
• an image processing step of performing at least one of the following processes for image processing of (i) a characteristic amount computation process of computing the amount of characteristic of the Bayer image based on the subimages obtained through frequency decomposition in the second frequency decomposing step, (ii) a correcting process on one or more of the subimages obtained through the frequency decomposition, and (iii) a set of information compression and information decompression processes on one or more of the subimages obtained through the frequency decomposition, in such a manner that processing details of the at least one of the processes for the high frequency component are different from those for the low frequency component;
• and
• first and second decomposed image combining steps of combining the subimages after the first and second frequency decomposing steps and the image processing step, to regenerate a single Bayer image.

A non-transitory computer-readable recording medium recording thereon a computer program causing a computer to execute the following steps of:

• an image obtaining step of obtaining a Bayer image in a Bayer array constituted by first rows in each of which red and green pixels are alternately arranged in a horizontal direction and second rows in each of which green and blue pixels are alternately arranged in the horizontal direction, the first rows and the second rows being alternately arranged in a vertical direction;
• a frequency decomposing step of decomposing the Bayer image, with respect to each of the vertical direction and the horizontal direction, into a low frequency component representing brightness information and a high frequency component representing color information, thereby to generate a plurality of subimages;
• an image processing step of performing at least one of the following processes for image processing of (i) a characteristic amount computation process of computing the amount of characteristic of the Bayer image based on the subimages obtained through frequency decomposition in the frequency decomposing step, (ii) a correcting process on one or more of the subimages obtained through the frequency decomposition, and (iii) a set of information compression and information decompression processes on one or more of the subimages obtained through the frequency decomposition, in such a manner that processing details of the at least one of the processes for the high frequency component are different from those for the low frequency component;
• and
• a decomposed image combining step of combining the subimages after the image processing step, to regenerate a single Bayer image, wherein
• the image processing step comprises:
• a filter obtaining step of obtaining first and second coefficient groups including a plurality of filter coefficients by which pixel values of pixels constituting each subimage before processed in the at least one of the processes are multiplied;
• and
• a filtering processing step of filtering the low frequency component using the first coefficient group obtained in the filter obtaining step and filtering the high frequency component using the second coefficient group obtained in the filter obtaining step,
• the first coefficient group is obtained based on a first weight monotonously decreasing depending on a distance from a current pixel, and
• the second coefficient group is obtained based on a second weight monotonously decreasing depending on the distance from the current pixel at a decrease rate smaller than that of the first weight, or having no change depending on the distance from the current pixel.