Image processing device, image processing system, and non-transitory computer readable medium
There is provided an image processing device. Pixels serving as a candidate for a nucleus are extracted from pixels included in a captured image obtained by imaging a sample including a target cell having the nucleus. A connected pixel group constituted of adjacent connected pixels is extracted from the extracted pixels. A value indicating a possibility that the target cell is included within a region surrounding the connected pixel group is determined. The value indicates the possibility being determined based on a value obtained by aggregating image feature amounts expressing luminance gradients determined based on luminance distribution with respect to a blue component within different partial regions included in the region surrounding the connected pixel group without performing normalization on each partial region.
1. An image processing device comprising:
a pixel extracting unit that extracts pixels serving as a candidate for a nucleus from pixels included in a captured image obtained by imaging a sample including a target cell having the nucleus;
a connected-pixel-group extracting unit that extracts a connected pixel group constituted of a predetermined number of adjacent connected pixels or more from the pixels extracted by the pixel extracting unit; and
a determining unit that determines a value indicating a possibility that the target cell is included within a region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit, the value indicating the possibility being determined based on a value obtained by aggregating image feature amounts expressing luminance gradients determined based on luminance distribution with respect to a blue component within a plurality of different partial regions included in the region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit without performing normalization on each partial region, the value indicating the possibility being determined also based on a condition for an image feature amount machine-learned based on a value obtained by aggregating image feature amounts expressing luminance gradients determined based on luminance distribution with respect to a blue component within partial regions included in a region occupied by a sample image without performing normalization on each partial region.
2. An image processing system comprising:
an image processing device;
an optical microscope that is connected to the image processing device; and
a display device that is connected to the image processing device,
wherein the optical microscope captures an image of a sample including a target cell having a nucleus,
wherein the image processing device includes
a pixel extracting unit that extracts pixels serving as a candidate for the nucleus from pixels included in the captured image obtained as a result of the optical microscope imaging the sample including the target cell having the nucleus,
a connected-pixel-group extracting unit that extracts a connected pixel group constituted of a predetermined number of adjacent connected pixels or more from the pixels extracted by the pixel extracting unit, and
a determining unit that determines a value indicating a possibility that the target cell is included within a region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit, the value indicating the possibility being determined based on a value obtained by aggregating image feature amounts expressing luminance gradients determined based on luminance distribution with respect to a blue component within a plurality of different partial regions included in the region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit without performing normalization on each partial region, the value indicating the possibility being determined also based on a condition for an image feature amount machine-learned based on a value obtained by aggregating image feature amounts expressing luminance gradients determined based on luminance distribution with respect to a blue component within partial regions included in a region occupied by a sample image without performing normalization on each partial region, and
wherein the display device displays an image expressing a position of the target cell in the captured image, the image being generated based on the value determined by the determining unit.
3. A non-transitory computer readable medium storing a program for causing a computer to function as:
a pixel extracting unit that extracts pixels serving as a candidate for a nucleus from pixels included in a captured image obtained by imaging a sample including a target cell having the nucleus;
a connected-pixel-group extracting unit that extracts a connected pixel group constituted of a predetermined number of adjacent connected pixels or more from the pixels extracted by the pixel extracting unit; and
a determining unit that determines a value indicating a possibility that the target cell is included within a region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit, the value indicating the possibility being determined based on a value obtained by aggregating image feature amounts expressing luminance gradients determined based on luminance distribution with respect to a blue component within a plurality of different partial regions included in the region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit without performing normalization on each partial region, the value indicating the possibility being determined also based on a condition for an image feature amount machine-learned based on a value obtained by aggregating image feature amounts expressing luminance gradients determined based on luminance distribution with respect to a blue component within partial regions included in a region occupied by a sample image without performing normalization on each partial region.
4. An image processing device comprising:
a captured-image acquiring unit that acquires a captured image obtained by imaging a sample including a target cell having a nucleus;
a pixel extracting unit that extracts pixels serving as a candidate for the nucleus from pixels included in the captured image;
a connected-pixel-group extracting unit that extracts a connected pixel group constituted of a predetermined number of adjacent connected pixels or more from the pixels extracted by the pixel extracting unit; and
a determining unit that determines a value indicating a possibility that the target cell is included within a region surrounding at least a portion of the connected pixel group extracted by the connected-pixel-group extracting unit, the value indicating the possibility being determined based on an image feature amount determined based on luminance distribution with respect to a plurality of different color components within the region, the value indicating the possibility being determined also based on a condition for an image feature amount machine-learned based on luminance distribution with respect to a plurality of different color components within a sample image of the target cell.
5. The image processing device according to claim 4 ,
wherein the determining unit determines the value indicating the possibility that the target cell is included within the region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit based on an image feature amount expressing a directed luminance gradient determined based on luminance distribution within the region and also based on a condition for an image feature amount expressing a directed luminance gradient machine-learned based on luminance distribution within the sample image of the target cell.
6. The image processing device according to claim 4 ,
wherein the value indicating the possibility that the target cell is included within the region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit is determined by the determining unit based on a value obtained by aggregating image feature amounts expressing luminance gradients determined based on luminance distribution within a plurality of different partial regions included in the region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit without performing normalization on each partial region, the value indicating the possibility being determined by the determining unit also based on a condition for an image feature amount machine-learned based on a value obtained by aggregating image feature amounts expressing luminance gradients determined based on luminance distribution within partial regions included in a region occupied by the sample image without performing normalization on each partial region.
7. The image processing device according to claim 4 ,
wherein the value indicating the possibility that the target cell is included within the region surrounding the connected pixel group extracted by the connected-pixel-group extracting unit is determined by the determining unit based on image feature amounts determined based on luminance distribution with respect to green and blue components within the region, the value indicating the possibility being determined by the determining unit also based on a condition for an image feature amount machine-learned based on luminance distribution with respect to green and blue components within the sample image of the target cell.
8. An image processing system comprising:
an image processing device;
an optical microscope that is connected to the image processing device; and
a display device that is connected to the image processing device,
wherein the optical microscope captures an image of a sample including a target cell having a nucleus,
wherein the image processing device includes
a captured-image acquiring unit that acquires the captured image obtained as a result of the optical microscope imaging the sample including the target cell having the nucleus,
a pixel extracting unit that extracts pixels serving as a candidate for the nucleus from pixels included in the captured image,
a connected-pixel-group extracting unit that extracts a connected pixel group constituted of a predetermined number of adjacent connected pixels or more from the pixels extracted by the pixel extracting unit, and
a determining unit that determines a value indicating a possibility that the target cell is included within a region surrounding at least a portion of the connected pixel group extracted by the connected-pixel-group extracting unit, the value indicating the possibility being determined based on an image feature amount determined based on luminance distribution with respect to a plurality of different color components within the region, the value indicating the possibility being determined also based on a condition for an image feature amount machine-learned based on luminance distribution with respect to a plurality of different color components within a sample image of the target cell.
9. A non-transitory computer readable medium storing a program for causing a computer to function as:
a captured-image acquiring unit that acquires a captured image obtained by imaging a sample including a target cell having a nucleus;
a pixel extracting unit that extracts pixels serving as a candidate for the nucleus from pixels included in the captured image;
a connected-pixel-group extracting unit that extracts a connected pixel group constituted of a predetermined number of adjacent connected pixels or more from the pixels extracted by the pixel extracting unit; and
a determining unit that determines a value indicating a possibility that the target cell is included within a region surrounding at least a portion of the connected pixel group extracted by the connected-pixel-group extracting unit, the value indicating the possibility being determined based on an image feature amount determined based on luminance distribution with respect to a plurality of different color components within the region, the value indicating the possibility being determined also based on a condition for an image feature amount machine-learned based on luminance distribution with respect to a plurality of different color components within a sample image of the target cell.