Learning data generation device, learning data generation method and recording medium
In a learning data generation device, an acquisition means acquires a first image corresponding to an image obtained by photographing one or more objects arranged in a predetermined arrangement state, and first information corresponding to information indicating the predetermined arrangement state in the first image. An image deformation means deforms a second image corresponding to an image of a desired object. An image generation means generates a fourth image having an arrangement state that matches or is similar to the predetermined arrangement state indicated by the first image, by pasting a deformed second image to the first image or a third image corresponding to the image of a position where the desired object is arranged. A learning data generation means generates data including the fourth image and second information corresponding to the information indicating the arrangement state of the object in the fourth image, as learning data.
1 . A learning data generation device comprising:
a memory configured to store instructions; and
one or more processors configured to execute the instructions to:
acquire a first image corresponding to an image obtained by photographing one or more objects arranged in a predetermined arrangement state, and first information corresponding to information indicating the predetermined arrangement state in the first image;
deform a second image corresponding to an image of a desired object;
generate a fourth image having an arrangement state that matches or is similar to the predetermined arrangement state indicated by the first image, by pasting a deformed second image to the first image or a third image corresponding to the image of a position where the desired object is arranged; and
generate data including the fourth image and second information corresponding to the information indicating the arrangement state of the object in the fourth image, as learning data.
2 . The learning data generation device according to claim 1 ,
wherein the one or more processors acquire, as the first information, information indicating one or more rectangular areas of a shape to surround a periphery of the object arranged in the predetermined arrangement state,
wherein the one or more processors deform the second image so as to fit within a deformed rectangular area which is deformed to conform to a size of the third image, and
wherein the one or more processors generate the fourth image by pasting the deformed second image to the third image.
3 . The learning data generation device according to claim 2 , wherein the one or more processors deform the second image while maintaining an aspect ratio of the second image.
4 . The learning data generation device according to claim 3 , wherein the one or more processors deform the second image using the aspect ratio set based on a length of a short side or a long side of the deformed rectangular area.
5 . The learning data generation device according to claim 2 ,
wherein the one or more processors acquire the second image of a first aspect ratio that matches or is similar to the aspect ratio of the deformed rectangular area, instead of deforming the second image, and
wherein the one or more processors generate the fourth image by pasting the second image of the first aspect ratio to the third image.
6 . The learning data generation device according to claim 2 ,
wherein the second image includes a fifth image which is an image similar to the deformed second image already pasted at the time of generating the fourth image,
wherein the one or more processors deform the fifth image so as to fit within the deformed rectangular area, and
wherein the one or more processors generate the fourth image by pasting a deformed fifth image at a position adjacent to the deformed second image already pasted at the time of generating the fourth image.
7 . The learning data generation device according to claim 1 ,
wherein the one or more processors acquire, as the first information, information indicating one or plural rectangular areas of a shape to surround a periphery of the object arranged in the predetermined arrangement state, and
wherein the one or more processors generate the fourth image by pasting the deformed second image to an area other than the rectangular area in the first image.
8 . The learning data generation device according to claim 1 , wherein the second image includes an image of a real object and an image of an unreal object.
9 . A learning data generation method comprising:
acquiring a first image corresponding to an image obtained by photographing one or more objects arranged in a predetermined arrangement state, and first information corresponding to information indicating the predetermined arrangement state in the first image;
deforming a second image corresponding to an image of a desired object;
generating a fourth image having an arrangement state that matches or is similar to the predetermined arrangement state indicated by the first image, by pasting a deformed second image to the first image or a third image corresponding to the image of a position where the desired object is arranged; and
generating data including the fourth image and second information corresponding to the information indicating the arrangement state of the object in the fourth image, as learning data.
10 . A non-transitory computer-readable recording medium recording a program, the program causing a computer to a processing comprising:
acquiring a first image corresponding to an image obtained by photographing one or more objects arranged in a predetermined arrangement state, and first information corresponding to information indicating the predetermined arrangement state in the first image;
deforming a second image corresponding to an image of a desired object;
generating a fourth image having an arrangement state that matches or is similar to the predetermined arrangement state indicated by the first image, by pasting a deformed second image to the first image or a third image corresponding to the image of a position where the desired object is arranged; and
generating data including the fourth image and second information corresponding to the information indicating the arrangement state of the object in the fourth image, as learning data.