Patent attributes
An image processing device including a storage unit configured to store an object detection neural network trained using a first K-channel image generated from a first M-channel image and a first N-channel image generated from the first M-channel image, a reception unit configured to receive, from a sensor, a second M-channel image and a second N-channel image that include an identical subject, and an image analysis unit configured to generate, using the object detection neural network trained using the first K-channel image, object detection result information with respect to a second K-channel image generated from the second M-channel image and the second N-channel image, and output the object detection result information.