Electronic device and solid-state imaging device
View Patent ↗The present invention suppresses an increase in processing time and power consumption that is associated with function implementation. An electronic device according to an embodiment comprises: an imaging unit ( 11 ) that generates image data; a processing unit ( 14 ) that executes neural network calculation model-based processing with respect to image data-based data that is read out from the imaging unit; a function execution unit ( 12 ) that performs a predetermined function on the basis of the results of the processing; and a detection unit ( 32 ) that detects displacement, wherein the processing unit executes the processing when the detection unit detects displacement.
1. An electronic device, comprising:
an image sensor that generates image data;
a processor configured to
execute neural network calculation model-based processing with respect to image data-based data that is read out from the image sensor;
execute processing utilizing a convolution neural network (CNN) with respect to data of an initially input region of the image data-based data of an identical frame; and
when the processing utilizing the CNN with respect to the data of the initially input region fails, processor is configured to execute processing utilizing a recurrent neural network (RNN) with respect to data of a unit region which is input next of the image data-based data of the identical frame;
a function execution circuit configured to perform a predetermined function to improve the image data on the basis of the results of the processing; and
a detection circuit configured to detect displacement of the electronic device, wherein the processor is configured to execute processing when the detection circuit detects any displacement.
2. An electronic device, comprising:
an image sensor that generates image data;
a processor configured to execute neural network calculation model-based processing with respect to image data-based data that is read out from the image sensor;
a function execution circuit configured to perform a predetermined function to improve the image data on the basis of the results of the processing; and
a detection circuit configured to detect displacement of the electronic device,
wherein the processor is configured to execute processing when the detection circuit detects any displacement,
wherein, during reproduction of the content by the application software, the processor is configured to recognize a facial expression using the neural network calculation model with respect to the image data-based data read out from the image sensor, and
wherein the function execution circuit is configured to use the social button to input a rating for the content according to the facial expression recognized by the processor.
3. An electronic device, comprising:
an image sensor that generates image data;
a processor configured to execute neural network calculation model-based processing with respect to image data-based data that is read out from the image sensor;
a function execution circuit configured to perform a predetermined function to improve the image data on the basis of the results of the processing; and
a detection circuit configured to detect displacement of the electronic device,
wherein the processor is configured to execute processing when the detection circuit detects any displacement,
wherein the processor is configured to perform first avoidance-target recognition using the neural network calculation model with respect to the image data-based data read out from the image sensor, and
wherein, when an avoidance target that is present ahead in the direction of advancement is recognized through the first avoidance-target recognition by the processor, the function execution circuit is configured to perform second avoidance-target recognition with respect to the image data-based data read out from the image sensor, and when the avoidance target that is present ahead in the direction of advancement is recognized once again through the second avoidance-target recognition, the function execution circuit is configured to issue an instruction externally to perform an operation to avoid the avoidance target.
4. An image processing device, comprising:
a processor configured to
execute processing using a neural network calculation model with respect to the image data-based data read out from an image sensor, wherein the processing includes
utilizing a convolution neural network (CNN) with respect to data of an initially input region of the image data-based data of an identical frame, and
on condition that processing utilizing the CNN with respect to the data of the initially input region fails, utilizing a recurrent neural network (RNN) with respect to data of a unit region which is input next of the image data-based data of the identical frame; and
output the results of the processing.