Neural network device and control method
AssigneeSony Group Corporation
InventorsSusumu Takatsuka, Kenta Endo, Christopher Wright, Bernadette Elliott-Bowman, Nicholas Walker
Life of this patent
Estimating the term…
Abstract
A neural network device includes a light receiving portion that receives light through an optical neural network, and a control unit that performs a predetermined control with detection of a predetermined target object based on a signal corresponding to light received by the light receiving portion as a trigger.
Claims14
A neural network device, comprising:
- a light receiving portion configured to:
- receive first light through a first optical neural network of a plurality of optical neural networks;
- and
- receive second light through a second optical neural network of the plurality of optical neural networks;
- a control unit configured to:
- detect a target object based on a first signal, wherein
- the first signal corresponds to the received first light, and
- the received first light corresponds to a trigger;
- and
- detect a region of interest based on a second signal that corresponds to the second light.
A control method, comprising
- in a neural network device:
- receiving first light through a first optical neural network of a plurality of optical neural networks;
- receiving second light through a second optical neural network of the plurality of optical neural networks;
- detecting target object based on a first signal, wherein
- the first signal corresponds to the received first light, and
- the received first light corresponds to a trigger;
- and
- detecting a region of interest based on a second signal that corresponds to the second light.
A non-transitory computer-readable medium having stored thereon, computer-executable instructions which when executed by a neural network device, cause the neural network device to execute operations, the operations comprising:
- receiving first light through a first optical neural network of a plurality of optical neural networks;
- receiving second light through a second optical neural network of the plurality of optical neural networks;
- detecting target object based on a first signal, wherein
- the first signal corresponds to the received first light, and
- the received first light corresponds to a trigger;
- and
- detecting a region of interest based on a second signal that corresponds to the second light.
Dependencies are inferred from the claim text.