Back to results
US 12,597,277 B2Granted
Learning method, learning device, mobile object control device, mobile object control method, and storage medium
AssigneeHonda Motor Co., Ltd.
InventorsYuji Yasui, Hideki Matsunaga, Takashi Matsumoto, Gakuyo Fujimoto
Life of this patent
Estimating the term…
Abstract
A learning method includes steps of associating, with an extended area in a bird's-eye view image, an annotation indicating that the extended area is a three-dimensional object; and generating, based on training data in which a bird's-eye view image is associated with an annotated bird's-eye view image obtained by assigning an annotation to a three-dimensional object in the bird's-eye view image, a trained model by learning parameters of a machine learning model so that the trained model receives input of a bird's-eye view image to output a three-dimensional object in the bird's-eye view image.
Claims12
A learning method executed by a computer comprising a processor and a memory, the learning method comprising:
- associating, with an extended area in a bird's-eye view image, an annotation indicating that the extended area is a three-dimensional object, the extended area being extended due to conversion of a camera image capturing the area into the bird's-eye view image;
- generating, based on training data in which a bird's-eye view image is associated with an annotated bird's-eye view image obtained by assigning an annotation to a three-dimensional object in the bird's-eye view image, a trained model by learning parameters of a machine learning model so that the trained model receives input of a bird's-eye view image and outputs a three-dimensional object in the bird's-eye view image;
- and
- causing a movement of a mobile object based on transmitting the trained model to the mobile object, wherein the trained model is stored in a storage of the mobile object.
1 more element
A learning device comprising:
- a non-transitory computer-readable storage medium storing computer-readable instructions;
- and
- a processor connected to the non-transitory computer-readable storage medium,
- wherein the processor executes the computer-readable instructions to:
- receive association of an extended area in a bird's-eye view image with an annotation indicating that the extended area is a three-dimensional object, the extended area being extended due to conversion of a camera image capturing the area into the bird's-eye view image;
- generate, based on training data in which a bird's-eye view image is associated with an annotated bird's-eye view image obtained by assigning an annotation to a three-dimensional object in the bird's-eye view image, a trained model by learning parameters of a machine learning model so that the trained model receives input of a bird's-eye view image and outputs a three-dimensional object in the bird's-eye view image;
- and
- cause a movement of a mobile object based on transmitting the trained model to the mobile object, wherein the trained model is stored in a storage of the mobile object.
5 more elements
A non-transitory computer readable storage medium storing a program thereon, the program causing a computer to:
- receive association of an extended area in a bird's-eye view image with an annotation indicating that the extended area is a three-dimensional object, the extended area being extended due to conversion of a camera image capturing the area into the bird's-eye view image,
- generate, based on training data in which a bird's-eye view image is associated with an annotated bird's-eye view image obtained by assigning an annotation to a three-dimensional object in the bird's-eye view image, a trained model by learning parameters of a machine learning model so that the trained model receives input of a bird's-eye view image and outputs a three-dimensional object in the bird's-eye view image;
- and
- cause a movement of a mobile object based on transmitting the trained model to the mobile object, wherein the trained model is stored in a storage of the mobile object.
1 more element
A learning method executed by a computer comprising a processor and a memory, the learning method comprising:
- associating, with an extended area in a bird's-eye view image, an annotation indicating that the extended area is a three-dimensional object, the extended area being extended due to conversion of a camera image capturing the area into the bird's-eye view image;
- generating, based on training data in which an image and a bird's-eye view image obtained by converting the image into a bird's-eye view coordinate system are associated with an image obtained by performing an inverse bird's-eye view conversion process for an annotated bird's-eye view image obtained by assigning an annotation to a three-dimensional object in the bird's-eye view image, a trained model by learning parameters of a machine learning model so that the trained model receives input of an image and a bird's-eye view image obtained by converting the image into a bird's-eye view coordinate system and outputs a three-dimensional object in the image;
- and
- causing a movement of a mobile object based on transmitting the trained model to the mobile object, wherein the trained model is stored in a storage of the mobile object.
1 more element
A learning device comprising:
- a non-transitory computer-readable storage medium storing computer-readable instructions;
- and
- a processor connected to the non-transitory computer-readable storage medium,
- wherein the processor executes the computer-readable instructions to:
- receive association of an extended area in a bird's-eye view image with an annotation indicating that the extended area is a three-dimensional object, the extended area being extended due to conversion of a camera image capturing the area into the bird's-eye view image;
- generate, based on training data in which an image and a bird's-eye view image obtained by converting the image into a bird's-eye view coordinate system are associated with an image obtained by performing an inverse bird's-eye view conversion process for an annotated bird's-eye view image obtained by assigning an annotation to a three-dimensional object in the bird's-eye view image, a trained model by learning parameters of a machine learning model so that the trained model receives input of an image and a bird's-eye view image obtained by converting the image into a bird's-eye view coordinate system and outputs a three-dimensional object in the image;
- and
- cause a movement of a mobile object based on transmitting the trained model to the mobile object, wherein the trained model is stored in a storage of the mobile object.
5 more elements
A non-transitory computer-readable storage medium storing a program thereon, the program causing a computer to:
- receive association of an extended area in a bird's-eye view image with an annotation indicating that the extended area is a three-dimensional object, the extended area being extended due to conversion of a camera image capturing the area into the bird's-eye view image;
- generate, based on training data in which an image and a bird's-eye view image obtained by converting the image into a bird's-eye view coordinate system are associated with an image obtained by performing an inverse bird's-eye view conversion process for an annotated bird's-eye view image obtained by assigning an annotation to a three-dimensional object in the bird's-eye view image, a trained model by learning parameters of a machine learning model so that the trained model receives input of an image and a bird's-eye view image obtained by converting the image into a bird's-eye view coordinate system and outputs a three-dimensional object in the image;
- and
- cause a movement of a mobile object based on transmitting the trained model to the mobile object, wherein the trained model is stored in a storage of the mobile object.
1 more element
Dependencies are inferred from the claim text.