Video management
AssigneeComcast Cable Communications, LLC
InventorSrinath V. Ramaswamy
Life of this patent
Estimating the term…
Abstract
The disclosure relates to a method of processing a sequence of image frames to reduce its length. One implementation may involve extracting coefficients (e.g., Discrete Cosine Transform coefficients) from components of individual frames, and comparing the resulting coefficients for sequential frames to identify frames having the least change from a prior frame. Also, scene change values for each frame may be calculated and placed in a sorted list to facilitate identification of frames for removal. Frame removal may be conducted in rounds, where a group of pictures (GOP) may only have one frame removed for any given round.
Claims20
A method, comprising:
- obtaining luminance component coefficient values for a plurality of image frames in a sequence of image frames;
- generating a scene change value for each of the plurality of image frames by comparing the coefficient values for sequential ones of the plurality of image frames;
- ranking the plurality of image frames according to their scene change values;
- and
- selecting one or more of the image frames for removal based on the ranked plurality of image frames.
A method, comprising:
- determining that a sequence of image frames is to be shortened;
- and
- selecting image frames of the sequence of image frames for removal based on comparisons between Discrete Cosine Transform (DCT) DC coefficients of sequential image frames in the sequence of image frames.
A method, comprising:
- determining a scene change value for each of a plurality of image frames in a sequence of image frames according to the following formula:
- C
-
- (
- v
- )
- =
- ∑
- n
- =
- 1
- n
- =
- h
-
- ∑
- m
- =
- 1
- m
- =
- w
-
- abs
-
- [
- DCv
-
- (
- m
- ,
- n
- )
- -
- DCu
-
- (
- m
- ,
- n
- )
- ]
- wherein
- C(v)=a scene change value for a vth image frame,
- DCv(m,n)=a Discrete Cosine Transform DC component of an m,n 8×8 block of DCT coefficients in the vth image frame,
- DCu(m,n)=a Discrete Cosine Transform DC component of an m,n 8×8 block of DCT coefficients in a uth image frame,
- v=u+1,
- w=a width of an image frame, and
- h=a height of an image frame;
- and
- using the scene change values for the plurality of image frames to select image frames for removal from the sequence of image frames.
Dependencies are inferred from the claim text.