Object pose estimation and tracking using machine learning

A system comprising:

• a processor;
• and
• a non-transitory computer-readable medium having stored thereon instructions that, when executed by the processor, cause the processor to perform operations comprising:
• receiving a video comprising a plurality of images representing an object;
• determining, using a machine learning model, based on a first image of the plurality of images, and for each respective vertex of a plurality of vertices of a bounding volume for the object, first corresponding two-dimensional coordinates of the respective vertex within the first image;
• tracking, from the first image to a second image of the plurality of images, a position of each respective vertex along a plane underlying the bounding volume;
• determining, for each respective vertex, second corresponding two-dimensional coordinates of the respective vertex within the second image based on the position of the respective vertex along the plane and independently of processing of the second image by the machine learning model;
• and
• determining, for each respective vertex, (i) first corresponding three-dimensional coordinates of the respective vertex, as represented in the first image, based on the first corresponding two-dimensional coordinates and (ii) corresponding second three-dimensional coordinates of the respective vertex, as represented in the second image, based on the second corresponding two-dimensional coordinates.

An article of manufacture including a non-transitory computer-readable medium, having stored thereon program instructions that, upon execution by a computing system, cause the computing system to perform operations comprising:

• receiving a video comprising a plurality of images representing an object;
• determining, using a machine learning model, based on a first image of the plurality of images, and for each respective vertex of a plurality of vertices of a bounding volume for the object, first corresponding two-dimensional coordinates of the respective vertex within the first image;
• tracking, from the first image to a second image of the plurality of images, a position of each respective vertex along a plane underlying the bounding volume;
• determining, for each respective vertex, second corresponding two-dimensional coordinates of the respective vertex within the second image based on the position of the respective vertex along the plane and independently of processing of the second image by the machine learning model;
• and
• determining, for each respective vertex, (i) first corresponding three-dimensional coordinates of the respective vertex, as represented in the first image, based on the first corresponding two-dimensional coordinates and (ii) corresponding second three-dimensional coordinates of the respective vertex, as represented in the second image, based on the second corresponding two-dimensional coordinates.