Techniques are discussed for providing mechanisms for coding and transmitting high definition video, e.g., over low bandwidth connections. In particular, foreground-objects are identified as distinct from the background of a scene represented in a plurality of video frames received from a video source, such as a camera. In identifying foreground-objects, semantically significant and semantically insignificant movement (e.g., repetitive versus non-repetitive movement) is differentiated. Processing of the foreground-objects and background proceed at different update rates or frequencies.
1. A transmission system comprising a processor configured to:
receive a plurality of frames;
analyze the plurality of frames;
determine a current bandwidth available for transmission;
determine, based on the current bandwidth, at least one of a pixel movement threshold or a pixel size threshold, the pixel movement threshold being a distance between a position in a first frame of a particular image object and a position in a second frame of the particular image object wherein the distance corresponds to a first number of pixels and the pixel size threshold being a difference between a dimension in the first frame of the particular image object and a dimension in the second frame of the particular image object wherein the difference corresponds to a second number of pixels;
identify foreground-object portions that are distinct from a background portion;
transmit, for each of the plurality of frames, foreground-object information for an identified foreground-object portion based on the identified foreground-object portion satisfying the at least one of the pixel movement threshold or the pixel size threshold; and
selectively transmit background information for less than all of the plurality of frames, wherein the background information is transmitted at a periodic rate based on the current bandwidth.
2. The system of claim 1 , wherein the foreground-object portions include semantically significant moving objects.
3. The system of claim 1 wherein the processor is further configured to compare features of the particular image object with an identified object in a previously received frame, the features comprising position, speed, movement direction, color, and size.
4. A non-transitory computer readable medium encoded with one or more computer readable instructions that, when executed by a processor, cause the processor to:
analyze video data from a plurality of images;
determine a current bandwidth available for transmission;
determine, based on the current bandwidth, at least one of a pixel movement threshold or a pixel size threshold, the pixel movement threshold being a distance between a position in a first frame of a particular image object and a position in a second frame of the particular image object wherein the distance corresponds to a first number of pixels and the pixel size threshold being a difference between a dimension in the first frame of the particular image object and a dimension in the second frame of the particular image object wherein the difference corresponds to a second number of pixels;
identify the presence of foreground-object portions that are distinct from a background portion in the video data;
transmit first portions of the video data, of different images, associated with at least one identified foreground-object portion, based on the at least one identified foreground-object portion satisfying the at least one of the pixel movement threshold or the pixel size threshold, at a first update rate; and
transmit second portions of the video data, of different images, associated with the background portion at a second update rate, wherein the first update rate is greater than the second update rate.
5. The computer readable medium of claim 4 , wherein the instructions that cause the processor to identify the presence of foreground-object portions cause the processor to identify semantically significant movement using a Gaussian model.
6. The computer readable medium of claim 5 , wherein the instructions that cause the processor to identify the presence of foreground-object portions cause the processor to identify a region of semantically significant movement as a foreground-object portion.
7. The computer readable medium of claim 4 , wherein the first update rate is thirty frames per second.
8. A video data transmission system comprising:
a processor configured to:
execute a scene analyzer algorithm for analyzing video data from a plurality of images;
determine a current bandwidth available for transmission;
determine, based on the current bandwidth, at least one of a pixel movement or a pixel size threshold, the pixel movement threshold being a distance between a position in a first frame of a particular image object and a position in a second frame of the particular image object wherein the distance corresponds to a first number of pixels and the pixel size threshold being a difference between a dimension in the first frame of the particular image object and a dimension in the second frame of the particular image object wherein the difference corresponds to a second number of pixels;
identify the presence of foreground-object portions that are distinct from a background portion in the video data; and
transmit first portions of the video data, of different images, associated with at least one identified foreground-object portion, based on the at least one identified foreground-object portion satisfying the at least one of the pixel movement threshold or the pixel size threshold, at a first update rate; and to transmit second portions of the video data, of different images, associated with the background portion at a second update rate, wherein the first update rate is greater than the second update rate.
9. The system of claim 8 , wherein the processor is configured to identify semantically significant movement using a Gaussian model.
10. The system of claim 9 , wherein the processor is configured to identify a region of semantically significant movement as a foreground-object portion.
11. The system of claim 8 , wherein the first update rate is thirty frames per second.
12. The system of claim 8 wherein the processor is configured to identify semantically significant moving objects as the foreground-object portions.