Cell image analysis method
A cell image analysis method according to this invention includes a step of acquiring a cell image ( 10 ) including a cell ( 90 ); a step of inputting the cell image to a learned model ( 6 ) that has learned classification of the cell into one of two or more types; a step of acquiring an index value ( 20 ) indicating accuracy of the classification of the cell that is included in the cell image into one of two or more types based on an analysis result of each of pixels of the cell image output from the learned model; and a step of displaying the acquired index value.
1 . A cell image analysis method comprising:
a step of acquiring a cell image including a cell;
a step of inputting the cell image to a learned model that has learned classification of the cell into one of two or more types;
a step of acquiring an index value indicating accuracy of the classification of the cell that is included in the cell image into one of two or more types based on an analysis result of each of pixels of the cell image output from the learned model; and
a step of displaying the acquired index value,
wherein the learned model has been learned to output a probability value(s) that is/are an estimation value(s) of the classification for every pixel of the cell image as the analysis result; and
a representative value of the probability value(s) obtained based on the probability value(s) output for every pixel of the cell image by the learned model in the step of acquiring an index value.
2 . The cell image analysis method according to claim 1 , wherein
the cell image includes cultivated cell that is cultivated in a cultivation container;
the learned model is produced by leaning at least one of classification of the cell whether a focus of the cell image is correct when the cell image is captured, classification of the cell whether a coating material on the cultivation container of the cell is proper, and classification of the cell whether the number of cultivation days of the cell is proper; and
an index value(s) of at least one of classification of the cell whether a focus of the cell image is correct when the cell image is captured, classification of the cell whether a coating material on the cultivation container of the cell is proper, and classification of the cell whether the number of cultivation days of the cell is/are proper is acquired as the index value in the step of acquiring an index value.
3 . The cell image analysis method according to claim 1 , wherein
the learned model is produced by learning classification of the cell whether the cell is suitable for analysis whether the cell is a normal or abnormal cell; and
a value representing a suitability degree for analysis whether the cell that is included in the cell image is a normal or abnormal cell is acquired based on the probability value(s) as the index value in the step of acquiring an index value.
4 . The cell image analysis method according to claim 3 , wherein the learned model is produced by learning classification of the cell whether cells of a common type are suitable for analysis whether each cell is a normal or abnormal cell.
5 . The cell image analysis method according to claim 1 further comprising a step of acquiring a cell area that is an area of the cell included in the cell image, wherein
the representative value of the probability value(s) in the cell area is obtained as the representative value of the probability value(s) in the step of obtaining the representative value of the probability value(s).
6 . The cell image analysis method according to claim 1 , wherein a superimposed cell image that is generated by superimposing numerical data of the representative value of the probability values and a distribution of the probability values on the cell image is displayed in the step of displaying the representative value of the probability value(s).
7 . The cell image analysis method according to claim 6 , wherein a frequency distribution of the probability values is displayed together with the numerical data of the representative value of the probability values and the superimposed cell image in the step of displaying the representative value of the probability value(s).
8 . The cell image analysis method according to claim 1 , wherein an average value of the probability values is obtained as the representative value in the step of obtaining the representative value of the probability value(s).
9 . The cell image analysis method according to claim 1 further comprising a step of producing the learned model by using teacher cell images that are the cell images, and teacher correct images that are generated by adding the cell images with a label value relating to at least two imaging conditions corresponding to the classification or a label value relating to at least two cultivation conditions corresponding to the classification.
10 . The cell image analysis method according to claim 9 , wherein the learned model is produced by using the teacher correct images that are added with two types of label values corresponding to whether a focus of the cell image is correct when the cell image is captured as the label value relating to the imaging conditions, or at least two types of label values relating to coating materials on a cultivation container in which the cell is cultivated, and the number of cultivation days of the cell as the label value relating to the cultivation conditions in the step of producing the learned model.
11 . The cell image analysis method according to claim 1 further comprising a step of determining whether the index value is greater than a threshold.