The Deep Learning Reconstruction operation allows you to run an image-to-image operation us-ing a trained DL model on any dataset. It could be used for various tasks such as getting prob-ability map of a segmentation model, image denoising, image enhancement, etc.