IMAGE GENERATION BASED ON VIDEO Russian patent published in 2021 - IPC H04N13/366 H04N13/117 G06T19/00 

FIELD: image processing.

SUBSTANCE: invention relates to the image generation from video elements and, in particular, but not exclusively, to the representation image generation for a virtual reality environment from stored video elements that may arise as a result of recording video of real-world environments. A device containing drive (101) for storing route data for a set of routes in N-dimensional space is proposed, wherein each route from the set of routes is associated with a video element containing frames containing both an image and depth information. Input tool (105) receives an indication of the observer’s position, and selection tool (107) selects the first route from the set of routes in accordance with a selection criterion depending on a distance indicator depending on the indication of the observer’s position and positions of routes from the set of routes. Extraction tool (103, 109) extracts the first video element associated with the first route from video source (203). Imaging tool (111) generates at least one representation image to indicate the observer’s position from the first set of frames from the first video element. In this system, the selection criterion is shifted to the currently selected route relatively to other routes from the set of routes.

EFFECT: provision of improved functioning, increased flexibility, increased image generation capabilities corresponding to different positions and directions of observation.

15 cl, 8 dwg