Neural network training using stereovision images

Hi guys,

I'm about to train an object detector neural network using Matlab, and I have a specific question concerning my images.

I'll obtain my images from a stereovision camera (ZED from Stereolabs), which will give me 2 sequences of images from both right and left camera.

My question is: is there a way to include this information in the neural network training ?

To be more precise, can I make Matlab to train the neural using both images at the same time, to ensure that the object detected will be the same on both image ?

Thank you in advance

Answer 1

[Srivardhan Gadila]

Please refer to the following Multiple-Input Networks and define your architecture in such a way that the network can take images from both left and right camera as inputs.

You can also refer to the following for other possible workarounds : Multiple Input Single Output Segmentation using Deep Learning