Multi-task image classifier for classifying inherently ordered values
View Patent ↗In an approach for classifying an image containing a continuous characteristic using an image classifier, an image tagged with a single value is received, wherein the single value is from a set of ordered value. A processor sets at least two thresholds based on the set of ordered values. A processor derives at least two labels according to the single value tagged in the image and the at least two thresholds. A processor runs image through an image classifier. A processor receives at least two predictions from the image classifier based on the at least two labels. A processor adjusts the image classifier based on differences between the at least two predictions and the at least two labels.
1. A method comprising:
receiving, by one or more processors, an image tagged with a single value, wherein the single value is from a set of ordered values;
setting, by one or more processors, at least two thresholds within the set of ordered values;
deriving, by one or more processors, at least two labels for the image from the single value tagged in the image and the at least two thresholds, wherein the at least two labels each represent a different subset range of ordered values of the set of ordered values;
running, by one or more processors, the image through an image classifier, wherein the image classifier classifies the image based on the at least two labels and the at least two thresholds;
receiving, by one or more processors, at least two predictions from the image classifier; and
adjusting, by one or more processors, the image classifier based on differences between the at least two predictions and the at least two labels.
2. The method of claim 1 , wherein the set of ordered values define a continuous characteristic in the image.
3. The method of claim 1 , wherein:
a first label of the at least two labels is designed around a first threshold of the at least two thresholds; and
the first label is selected from the group consisting of a first label option denoting that the continuous characteristic being analyzed is at or above the first threshold and a second label option denoting that the continuous characteristic being analyzed is below the first threshold.
4. The method of claim 1 , wherein:
a second label of the at least two labels is designed around a second threshold of the at least two thresholds; and
the second label is selected from a group consisting of a first label option denoting that the continuous characteristic being analyzed is at or above the second threshold and a second label option denoting that the continuous characteristic being analyzed is below the second threshold.
5. The method of claim 1 , wherein a first prediction of the at least two predictions is selected from the group consisting of a positive prediction denoting that the first threshold was met and a negative prediction denoting that the first threshold was not met.
6. The method of claim 1 , wherein a second prediction of the at least two predictions is selected from the group consisting of a positive prediction denoting that the second threshold was met and a negative prediction denoting that the second threshold was not met.
7. A computer system comprising:
one or more computer processors;
one or more computer readable storage media;
program instructions stored on the computer readable storage media for execution by at least one or more processors, the program instructions comprising:
program instructions to receive an image tagged with a single value, wherein the single value is from a set of ordered values;
program instructions to set at least two thresholds within the set of ordered values;
program instructions to derive at least two labels for the image from the single value tagged in the image and the at least two thresholds, wherein the at least two labels each represent a different subset range of ordered values of the set of ordered values;
program instructions to run the image through an image classifier, wherein the image classifier classifies the image based on the at least two labels and the at least two thresholds;
program instructions to receive at least two predictions from the image classifier; and
program instructions to adjust the image classifier based on differences between the at least two predictions and the at least two labels.
8. The computer system of claim 7 , wherein the set of ordered values define a continuous characteristic in the image.
9. The computer system of claim 7 , wherein:
a first label of the at least two labels is designed around a first threshold of the at least two thresholds; and
the first label is selected from the group consisting of a first label option denoting that the continuous characteristic being analyzed is at or above the first threshold and a second label option denoting that the continuous characteristic being analyzed is below the first threshold.
10. The computer system of claim 7 , wherein:
a second label of the at least two labels is designed around a second threshold of the at least two thresholds; and
the second label is selected from a group consisting of a first label option denoting that the continuous characteristic being analyzed is at or above the second threshold and a second label option denoting that the continuous characteristic being analyzed is below the second threshold.
11. The computer system of claim 7 , wherein a first prediction of the at least two predictions is selected from the group consisting of a positive prediction denoting that the first threshold was met and a negative prediction denoting that the first threshold was not met.
12. The computer system of claim 7 , wherein a second prediction of the at least two predictions is selected from the group consisting of a positive prediction denoting that the second threshold was met and a negative prediction denoting that the second threshold was not met.
13. A computer program product comprising:
one or more computer readable storage media and program instructions stored on the one or more computer readable storage media, the program instructions comprising:
program instructions to receive an image tagged with a single value, wherein the single value is from a set of ordered values;
program instructions to set at least two thresholds within the set of ordered values;
program instructions to derive at least two labels for the image from the single value tagged in the image and the at least two thresholds, wherein the at least two labels each represent a different subset range of ordered values of the set of ordered values;
program instructions to run the image through an image classifier, wherein the image classifier classifies the image based on the at least two labels and the at least two thresholds;
program instructions to receive at least two predictions from the image classifier; and
program instructions to adjust the image classifier based on differences between the at least two predictions and the at least two labels.
14. The computer program product of claim 13 , wherein the set of ordered values define a continuous characteristic in the image.
15. The computer program product of claim 13 , wherein:
a first label of the at least two labels is designed around a first threshold of the at least two thresholds; and
the first label is selected from the group consisting of a first label option denoting that the continuous characteristic being analyzed is at or above the first threshold and a second label option denoting that the continuous characteristic being analyzed is below the first threshold.
16. The computer program product of claim 13 , wherein:
a second label of the at least two labels is designed around a second threshold of the at least two thresholds; and
the second label is selected from a group consisting of a first label option denoting that the continuous characteristic being analyzed is at or above the second threshold and a second label option denoting that the continuous characteristic being analyzed is below the second threshold.
17. The computer program product of claim 13 , wherein each of the at least two predictions is selected from the group consisting of a positive prediction denoting that the second threshold was met and a negative prediction denoting that the second threshold was not met.