Patch-based depth mapping method and apparatus for high-efficiency encoding/decoding of plenoptic video

A method performed by an electronic apparatus to perform encoding on a plenoptic video, the method comprising:

• generating an atlas including a patch by classifying a reference image and an additional image from the plenoptic video and then by using the additional image;
• generating an encoding code by encoding the reference image and the atlas using a 2D video codec;
• and
• storing information including the encoding code and metadata or generating the information as a bit-stream and transmitting it to other apparatus,
• wherein generating the atlas comprises:
• mapping a depth value by applying minimum-maximum linear scaling in units of an input viewpoint image;
• and
• finding a minimum value and a maximum value of the depth value per the patch and remapping the depth value through minimum-maximum linear scaling,
• wherein the finding comprises storing information of a minimum value and a maximum value of the depth value before mapping the depth value per the patch in the metadata.

An apparatus, comprising:

• a memory storing a plenoptic video;
• and
• a controller configured to control to classify a reference image and an additional image from the plenoptic video during encoding the plenoptic video and then generate an atlas including a patch using the additional image,
• wherein the controller is configured to:
• map a depth value by applying minimum-maximum linear scaling in units of an input viewpoint image,
• find a minimum value and a maximum value of the depth value per the patch and remapping the depth value through minimum-maximum linear scaling, and
• store information of a minimum value and a maximum value of the depth value before mapping the depth value per the patch in metadata when the controller finds the minimum value and the maximum value of the depth value.