Hello,
I’m planning to build a synthetic RGB-D image dataset for training neural networks. Unreal Engine can produce realistic pictures, which is an advantage for building synthetic datasets.
The dataset should consist of a large number of multi-view RGB-D images. An RGB-D image consists of a color image and a depth image. Thus I need to obtain both color images and depth images from the same camera. Also, I want to place multiple cameras in the scene and capture RGB-D images from different viewpoints. Since the number of images to be captured is large, I probably need to set up a dynamic scene and let the cameras capture videos (i.e., image sequences).
Can this be done using Unreal Engine? Thanks!