Data processing systems for real-time camera parameter estimation

A non-transitory computer-readable medium storing computer-executable instructions for:

• identifying semantic keypoints in an image;
• identifying person keypoints in the image;
• generating a first estimated homography using the semantic keypoints;
• projecting a location of each of the person keypoints onto a coordinate system;
• identifying, for each of the semantic keypoints, a semantic keypoint location in the coordinate system;
• identifying, for each of the person keypoints for the image, based at least in part on projecting the location of each of the person keypoints onto the coordinate system, a proximate person keypoint location in the coordinate system;
• iteratively generating one or more subsequent estimated homographies based at least in part on the semantic keypoints for the image, the semantic keypoint location for each of the semantic keypoints for the image, the person keypoints for the image, and the proximate person keypoint location for each of the person keypoints for the image;
• selecting a particular homography of the one or more subsequent estimated homographies as a final estimated homography;
• determining one or more intrinsic parameter estimates based at least in part on the final estimated homography;
• and
• determining one or more extrinsic parameter estimates based at least in part on the final estimated homography.

A computer-implemented data-processing method for camera parameter estimation, the method comprising:

• determining one or more semantic keypoints in a particular image of a video sequence;
• determining one or more person keypoints in the particular image;
• generating a first estimated homography using the one or more semantic keypoints;
• projecting each of the one or more person keypoints onto a coordinate system associated with a particular environment;
• determining, for each of the one or more semantic keypoints, a semantic keypoint location in the coordinate system;
• determining, for each of the one or more person keypoints, based at least in part on projecting each of the one or more person keypoints onto the coordinate system, a proximate person keypoint location in the coordinate system;
• generating one or more subsequent estimated homographies based at least in part on the one or more semantic keypoints, the semantic keypoint location for each of the one or more semantic keypoints, the one or more person keypoints, and the proximate person keypoint location for each of the one or more person keypoints;
• determining a final estimated homography based at least in part on the one or more subsequent estimated homographies;
• determining one or more intrinsic parameter estimates based at least in part on the final estimated homography;
• and
• determining one or more extrinsic parameter estimates based at least in part on the final estimated homography.