| CPC G06V 10/803 (2022.01) [G06V 10/26 (2022.01); G06V 10/82 (2022.01); G06V 20/52 (2022.01)] | 11 Claims |

|
1. A method for optimizing a neural network, comprising:
capturing, via a first sensor group having a first field of view, a first sample set having a first sensor domain corresponding to the first field of view;
capturing, via a second sensor group having a second field of view, a second sample set having a second sensor domain corresponding to the second field of view, wherein the second field of view overlaps at least a portion of the first field of view;
generating one or more regions of interest of the second sample set; translating the one or more regions of interest from the second sensor domain to the first sensor domain;
identifying one or more nodes of the neural network which correspond to the one or more translated regions of interest, wherein each node has a weight value; and
optimizing the neural network by at least one of:
increasing the weight value of the nodes corresponding to the one or more translated regions of interest; and
decreasing the weight value of the nodes not corresponding to the one or more translated regions of interest;
wherein the nodes of the neural network corresponding to the translated regions of interest are identified by:
feeding a binary input matrix into the neural network; and
detecting the nodes of the neural network which are activated by feeding the binary matrix; and
wherein the binary input matrix is equal in dimension to the first sample set; wherein the binary input matrix comprises one or more zero entries corresponding to the areas outside of the translated regions of interest; and wherein the binary input matrix comprises one or more non-zero entries corresponding to the translated regions of interest.
|