Systems and methods for utilizing a deep learning model to determine vehicle viewpoint estimations

A device may receive a first image . The device may process the first image to identify an object in the first image and a location of the object within the first image . The device may extract a second image from the first image based on the location of the object within the first image . The device may process the second image to determine at least one of a coarse - grained viewpoint estimate or a fine - grained view point estimate associated with the object . The device may determine an object viewpoint associated with the second vehicle based on the at least one of the coarse - grained viewpoint estimate or the fine - grained viewpoint estimate . The device may perform one or more actions based on the object viewpoint .