Face discontinuity filtering for 360-degree video coding

A method for decoding video data associated with a 360-degree video, the method comprising:

• obtaining a frame-packed picture from the video data, wherein the frame-packed picture comprises one or more faces that correspond to respective two-dimensional (2D) projections of the 360-degree video onto a 2D plane;
• determining whether a face discontinuity exists in the frame-packed picture;
• and
• based at least on a determination that the face discontinuity exists in the frame-packed picture:
• identifying a first set of samples and a second set of samples that contribute to the face discontinuity, wherein the first set of samples is associated with a first face among the one or more faces, and wherein the second set of samples is associated with a second face among the one or more faces;
• filtering the first set of samples based on first padded samples associated with the first face;
• and
• filtering the second set of samples based on second padded samples associated with the second face;
• wherein the first padded samples are located outside the boundaries of the first face and are derived based on sample locations inside the boundaries of the first face that bear a first geometrical relationship with the first padded samples, wherein the second padded samples are located outside the boundaries of the second face and are derived based on sample locations inside the boundaries of the second face that bear a second geometrical relationship with the second padded samples, and wherein the filtering of the first set of samples and the second set of samples removes one or more artifacts associated with the face discontinuity.

A video encoding device, comprising:

• a processor configured to:
• determine a frame-packed picture associated with a 360-degree video, wherein the frame-packed picture comprises one or more faces that correspond to respective two-dimensional (2D) projections of the 360-degree video onto a 2D plane;
• determine whether a face discontinuity exists in the frame-packed picture;
• and
• based on a determination that the face discontinuity exists in the frame-packed picture:
• identify a first set of samples and a second set of samples that contribute to the face discontinuity, wherein the first set of samples is associated with a first face among the one or more faces, and wherein the second set of samples is associated with a second face among the one or more faces;
• filter the first set of samples based on first padded samples associated with the first face;
• and
• filter the second set of samples based on second padded samples associated with the second face;
• wherein the first padded samples are located outside the boundaries of the first face and are derived based on sample locations inside the boundaries of the first face that bear a first geometrical relationship with the first padded samples, wherein the second padded samples are located outside the boundaries of the second face and are derived based on sample locations inside the boundaries of the second face that bear a second geometrical relationship with the second padded samples, and wherein the filtering of the first set of samples and the second set of samples removes one or more artifacts associated with the face discontinuity.