DEEP LEARNING BASED SELECTION OF SAMPLES FOR ADAPTIVE SUPERSAMPLING

摘要

An apparatus for facilitating deep learning-based sample selection for adaptive supersampling is disclosed. The device receives training data comprising an input tile and a corresponding supersampling value for the input tile, each input tile comprising a plurality of pixels, based on the training data, a rendered pixel tile. one or more processing elements for training the machine learning model to identify a supersampling level for