SIDE CONDITIONS OF SIMULTANEOUS POSITION DETERMINATION AND CARD PROCESSING IN GENERATIVE ADVERSARIAL NETWORKS FOR MONOCULAR DEPTH ESTIMATION

摘要

The present disclosure provides constraints for simultaneous position determination and map generation in Generative Adversarial Networks for monocular depth estimation. The disclosure relates to systems, methods, and apparatus for determining a depth map of an environment based on a monocular image. A method of determining a depth map involves receiving a plurality of images that form an image sequence from a monocular camera. The method includes determining pose vector data for two consecutive frames of the frame and providing the frame sequence and pose vector data to a Generative Adversarial Network (GAN), wherein the GAN is trained using temporal constraints to generate a depth map for each frame of the frame , The method includes generating a reconstructed image based on a depth map received from the GAN.