Image processing device, cell recognition device, cell recognition method, and cell recognition program

A cell recognition device comprising:

• an image acquisition device that is configured to acquire a cell image formed by capturing an image of a cell cluster composed of a plurality of cells;
• and
• an image processing device that includes a memory and a processor, the processor configured to:
• calculate a feature value, the feature value representing how likely a pixel value in each of pixels in the cell image acquired by the image acquisition device is to be an extreme value;
• detect, as peak positions, pixel positions the feature value of which are greater than a prescribed feature value threshold value in the cell image;
• record the detected peak positions in the memory;
• form, one at a time for the recorded peak positions recorded in the memory, a cell region on the basis of a distribution of the pixel values of a plurality of pixels included in a region peripheral to the peak position in the cell image;
• identify a center position of the formed cell region;
• determine, by using at least one of the peak position, a morphology of the cell region, and the center position of the cell region, a proximity state between the currently formed cell region and a previously formed cell region;
• and
• correct, when it is determined that the proximity state is satisfied, at least one of the currently formed cell region and the previously formed cell region.

A cell recognition method comprising:

• calculating a feature value, the feature value representing how likely a pixel value in each of pixels in a cell image formed by capturing an image of a cell cluster composed of a plurality of cells is to be an extreme value;
• detecting, as peak positions, pixel positions the feature value of which are greater than a prescribed feature value threshold value in the cell image and recording the detected peak positions;
• forming, one at a time for the recorded peak positions, a cell region on the basis of a distribution of the pixel values of a plurality of pixels included in a region peripheral to the peak position in the cell image;
• identifying a center position of the formed cell region;
• determining, by using at least one of the peak position, a morphology of the cell region, and the center position of the cell region, a proximity state between the currently formed cell region and a previously formed cell region;
• and
• correcting, when it is determined that the proximity state is satisfied, at least one of the currently formed cell region and the previously formed cell region.

A non-transitory computer-readable medium having a cell recognition program stored thereon, the cell recognition program causing a computer to execute functions of:

• calculating a feature value, the feature value representing how likely a pixel value in each of pixels in a cell image formed by capturing an image of a cell cluster composed of a plurality of cells is to be an extreme value;
• detecting, as peak positions, pixel positions the feature value of which are greater than a prescribed feature value threshold value in the cell image and recording the peak positions;
• forming, one at a time for the recorded peak positions, a cell region on the basis of a distribution of the pixel values of a plurality of pixels included in a region peripheral to the peak position in the cell image;
• identifying a center position of the formed cell region;
• determining, by using at least one of the peak position, a morphology of the cell region, and the center position of the cell region, a proximity state between the currently formed cell region and a previously formed cell region;
• and
• correcting, when it is determined that the proximity state is satisfied, at least one of the currently formed cell region and the previously formed cell region.