Information processing apparatus, information processing method, and recording medium
An information processing apparatus including: a classifier that is applied for medical treatment data; and a hardware processor that sets a property of the classifier, evaluates a sufficiency level of the property by using evaluation data for the classifier, and selects and sets one classifier as the classifier to be used in a facility by referring to an evaluation result.
1. An information processing apparatus comprising:
a classifier that is applied for medical treatment data; and
a hardware processor that:
sets a property of the classifier;
evaluates a sufficiency level of the property by using evaluation data for the classifier;
selects and sets one classifier as the classifier to be used in a facility by referring to an evaluation result; and
sets a learning condition of the classifier based on the set property of the classifier, wherein the hardware processor extracts learning data used in learning of the classifier in accordance with the property of the classifier so as to reduce a loss at the time of learning.
2. The information processing apparatus according to claim 1 , wherein the property includes both or one of a detection target to be detected and a detection performance for the detection target.
3. The information processing apparatus according to claim 2 , wherein the detection performance includes both or one of a sensitivity that is a probability of classifying data including an abnormality as abnormal and a specificity that is a probability of classifying normal data as normal.
4. The information processing apparatus according to claim 1 , wherein the classifier that is an evaluation target is a learned classifier to which learning is performed by using data collected in a present facility that is the facility to use the classifier.
5. The information processing apparatus according to claim 1 , wherein the classifier that is an evaluation target is a learned classifier to which learning is performed by using data collected in a facility different from the facility to use the classifier.
6. The information processing apparatus according to claim 1 , wherein the property includes both or one of a detection target to be detected and a detection performance for the detection target, wherein
the learning data that is used in the learning of the classifier is sorted based on the property of the classifier, and
the hardware processor extracts the learning data in accordance with the set property of the classifier.
7. The information processing apparatus according to claim 1 , further comprising a display that displays the property of the classifier and the evaluation result regarding the classifier.
8. The information processing apparatus according to claim 7 , wherein the display uses different displaying methods for the evaluation result that satisfies the property of the classifier set by the hardware processor and for the evaluation result that does not satisfy the set property of the classifier.
9. The information processing apparatus according to claim 1 , wherein
the hardware processor changes the learning condition and performs learning again when the hardware processor extracts the learning data in accordance with the set property of the classifier.
10. The information processing apparatus according to claim 1 , wherein
the loss at the time of learning is an evaluation value which is calculated from an output image of the classifier, the calculation using both or one of a maximum value and an area ratio.
11. The information processing apparatus according to claim 1 , wherein
the hardware processor extracts the learning data in accordance with the property of the classifier by changing weighting to a loss of a normal image from the weighting to the loss of an abnormal image.
12. The information processing apparatus according to claim 1 , wherein
the property includes a detection performance for a detection target, and
the detection performance includes both or one of a sensitivity that is a probability of classifying data including an abnormality as abnormal and a specificity that is a probability of classifying normal data as normal.
13. An information processing method, causing a hardware processor to:
set a property of a classifier applied for medical treatment data;
evaluate a sufficiency level of the property by using evaluation data for the classifier;
select and set one classifier as the classifier to be used in a facility by referring to an evaluation result of the evaluating; and
set a learning condition of the classifier based on the set property of the classifier, wherein the method further causes the hardware processor to
extract learning data used in learning of the classifier in accordance with the property of the classifier so as to reduce a loss at the time of learning.
14. A non-transitory recording medium storing a computer readable program causing a computer of an information processing apparatus to perform:
setting a property of a classifier applied for medical treatment data;
evaluating a sufficiency level of the property by using evaluation data for the classifier;
selecting and setting one classifier as the classifier to be used in a facility by referring to an evaluation result of the evaluating; and
setting a learning condition of the classifier based on the set property of the classifier, wherein the program further causes the computer to perform
extracting learning data used in learning of the classifier in accordance with the property of the classifier so as to reduce a loss at the time of learning.