Object detection device, object detection method, and recording medium
Provided is an object detection device or the like which efficiently generates good-quality training data. This object detection device is provided with: a detection unit which uses a dictionary to detect objects from an input image; a reception unit which displays, on a display device, the input image accompanied by a display emphasizing partial areas of detected objects, and receives, from one operation of an input device, a selection of a partial area and an input of the class of the selected partial area; a generation unit which generates training data from the image of the selected partial area and the inputted class; and a learning unit which uses the training data to learn the dictionary.
1. An object detection device comprising:
one or more processors acting as a detecting unit configured to detect an object from an input image by referring to a dictionary;
the one or more processors acting as an accepting unit configured to display, on a display device, the input image with indication that emphasizes a partial region of the detected object and accepting a selection of the partial region and an input of a class for the selected partial region through one operation of an input device;
the one or more processors acting as a generating unit configured to generate teacher data from an image of the selected partial region and the input class; and
the one or more processors acting as a learning unit configured to learn the dictionary based on the teacher data,
wherein the accepting unit accepts the selection of the partial region based on a location on the input image subjected to the one operation.
2. The object detection device according to claim 1 , wherein the accepting unit accepts the selection of the partial region that has, in the region, the location on the input image subjected to the one operation, and, if there is no partial region that has, in the region, the location on the input image subjected to the one operation, calculates a partial region of the object based on the location on the input image subjected to the one operation.
3. The object detection device according to claim 1 , wherein the accepting unit accepts an input of the class by a type of the one operation.
4. The object detection device according to claim 3 , wherein the accepting unit accepts an input of a preset first class when the type of the one operation is a first type, and accepts an input of a class that is incremented or decremented from a current class of the partial region when the type of the one operation is a second type.
5. The object detection device according to claim 1 , wherein:
the accepting unit accepts an input of a different class depending on whether there is the partial region that has, in the region, the location on the input image subjected to one operation.
6. The object detecting device according to claim 5 , wherein
the accepting unit accepts an input of a first class when there is the partial region that has, in the region, the location on the input image subjected to the one operation and accepts an input of a second class when there is no such partial region.
7. The object detecting device according to claim 5 , wherein
the accepting unit accepts an input of the first class for the existing partial portion when there is the partial region that has, in the region, the location on the input image subjected to one operation, as well as, calculates a partial region of the object based on the location on the input image subjected to the one operation and accepts an input of a second class for the calculated partial region.
8. The object detecting device according to claim 6 , wherein the accepting unit determines the second class by a type of the one operation.
9. The object detecting device according to claim 1 , wherein
the one operation of the input device is a single mouse gesture.
10. The object detecting device according to claim 1 , wherein
the one operation of the input device is a single touch gesture.
11. The object detecting device according to claim 1 , wherein
the accepting unit hides the indication that emphasizes the partial region of a detected object if the partial region of the detected object includes the same location and pre-input class as those of a partial region for which the class has previously been input.
12. An object detecting method comprising:
detecting an object from an input image by referring to a dictionary;
displaying, on a display device, the input image with indication that emphasizes a partial region of the detected object;
accepting a selection of the partial region and an input of a class for the selected partial region through one operation of an input device;
generating teacher data from an image of the selected partial region and the input class;
learning the dictionary based on the teacher data; and
accepting the selection of the partial region based on a location on the input image subjected to the one operation.
13. A non-transitory computer readable recording medium that stores a program for causing a computer to implement functions of:
detecting an object from an input image by referring to a dictionary;
displaying, on a display device, the input image with indication that emphasizes a partial region of the detected object;
accepting a selection of the partial region and an input of a class for the selected partial region through one operation of an input device;
generating teacher data from an image of the selected partial region and the input class;
learning the dictionary based on the teacher data; and
accepting the selection of the partial region based on a location on the input image subjected to the one operation.
14. The object detection device according to claim 1 , wherein the accepting unit accepts the selection of the partial region that has, in the region, the location on the input image subjected to the one operation, and, if there is no partial region that has, in the region, the location on the input image subjected to the one operation, calculates a partial region of the object based on the location on the input image subjected to the one operation.
15. The object detection device according to claim 1 , wherein the accepting unit accepts an input of the class by a type of the one operation.
16. The object detection device according to claim 2 , wherein the accepting unit accepts an input of the class by a type of the one operation.
17. The object detection device according to claim 1 , wherein:
the accepting unit accepts an input of a different class depending on whether there is the partial region that has, in the region, the location on the input image subjected to one operation.
18. The object detection device according to claim 2 , wherein:
the accepting unit accepts an input of a different class depending on whether there is the partial region that has, in the region, the location on the input image subjected to one operation.