Information processing device, image recognition method and non-transitory computer readable medium
According to one embodiment, an information processing device includes a camera and a processing circuit. The camera takes first distance images of an object for a plurality of angles. The processing circuit generates a three-dimensional model of the object based on the first distance image, and generates an extracted image indicating a specific region of the object corresponding to the plurality of angles based on the three-dimensional model.
1. An information processing device comprising:
a camera configured to generate first distance images of an object for a plurality of angles; and
a processing circuit configured to:
generate three-dimensional models of the object based on each of the first distance images;
generate extracted images indicating a specific region in the object from the three-dimensional models;
generate a first model which detects the object within an arbitrary second distance image, based on the first distance images; and
generate a second model which estimates the location of the specific region in the second distance image, based on the extracted images and the first distance images.
2. The information processing device according to claim 1 , wherein
the processing circuit determines the specific region of the object by searching the region of the object which satisfies a condition that includes at least either a relative location in the object, a flat area on the object or a width of an edge of the object, in the three-dimensional model.
3. The information processing device according to claim 1 , wherein
the processing circuit determines the specific region of the object, based on color information of the object.
4. The information processing device according to claim 1 , wherein
the processing circuit generates a new first distance image by performing at least either expansion operations, contraction operations or rotation operations to the plurality of first distance images, and generates a three-dimensional model of the object based on the new first distance image.
5. The information processing device according to claim 1 , wherein
the processing circuit generates a combination of the first model and the second model for a plurality of the objects.
6. The information processing device according to claim 5 , wherein
the processing circuit uses back propagation method in a neural network for at least either generation of the first model or generation of the second model.
7. The information processing device according to claim 1 , wherein
the processing circuit estimates the location of the specific region by applying the second model to the second distance image taken from an arbitrary angle.
8. The information processing device according to claim 1 , wherein
the processing circuit estimates the location of the specific region by applying the second model, when the object is detected by applying the first model to the second distance image taken from an arbitrary angle.
9. The information processing device according to claim 5 , wherein
the processing circuit estimates the location of the specific region by applying the second model, when the object is detected by applying either of the first models to the second distance image taken from an arbitrary angle, and
the second model which is applied by the processing circuit is the second model for estimating the location of the specific region in the object detected by the first model.
10. The information processing device according to claim 8 , wherein
the processing circuit determines the location a robotic hand grips as the specific region of the object, and provides locational information including the estimated location of the specific region in the object as control signals of the robotic hand.
11. An information processing device according to claim 8 , further comprising:
a display displaying at least either the first distance image, the three-dimensional model, the extracted image, the first model or the second model.
12. An image recognition method comprising:
generating first distance images of an object from a plurality of angles;
generating three-dimensional models of the object based on each of the first distance images;
generating extracted images indicating a specific region in the object from the three-dimensional models;
generating a first model which detects the object within an arbitrary second distance image based on the first distance images; and
generating a second model which estimates the location of the specific region in the second distance image, based on the extracted images and the first distance images.
13. A non-transitory computer readable medium having a computer program stored therein which causes a computer to execute processes comprising:
generating first distance images of an object from a plurality of angles;
generating three-dimensional models of the object based on each of the first distance images;
generating extracted images indicating a specific region in the object from the three-dimensional models;
generating a first model which detects the object within an arbitrary second distance image, based on the first distance images; and
generating a second model which estimates the location of the specific region in the second distance image, based on the extracted images and the first distance images.