Method, system and apparatus for segmenting an image of a scene
A method of segmenting an image of a scene captured using one of a plurality of cameras in a network. A mask of an image of a scene captured by a first one of said cameras is received. A set of pixels in the mask likely to be in a foreground of an image captured by a second one of said cameras is determined based on the received mask, calibration information, and a geometry of the scene. A set of background pixels for the second camera is generated based on the determined set of pixels. The generated set of background pixels is transmitted to the second camera. The image of the scene captured by the second camera is segmented using the transmitted background pixels.
1. A method of segmenting an image of a scene captured using one of a plurality of cameras in a network, the method comprising:
receiving a mask of an image of a scene captured by a first one of said cameras;
determining a set of pixels in the mask likely to be in a foreground of an image captured by a second one of said cameras based on the received mask, calibration information, and a geometry of the scene;
generating a set of background pixels for the second camera based on the determined set of pixels;
transmitting the generated set of background pixels to the second camera; and
segmenting the image of the scene captured by the second camera using the transmitted background pixels.
2. The method according to claim 1 , wherein generating the set of background pixels uses a further image of the scene captured by the first camera.
3. The method according to claim 1 , wherein the second camera is selected from the plurality of cameras based on similarity with respect to the first camera.
4. The method according to claim 3 , wherein similarity is further based on background pixels, similarity of views between the cameras.
5. The method according to claim 1 , wherein the generated set of background pixels is transmitted at lower resolution for one or more image frames.
6. The method according to claim 1 , wherein scene geometry includes a ground plane associated with the scene.
7. The method according to claim 1 , wherein the second camera is selected to transmit the generated set of background pixels based on a weight of the background pixels.
8. The method according to claim 1 , wherein the weight of the background pixels is based on a similarity measure.
9. A system for segmenting an image of a scene captured using one of a plurality of cameras in a network, the system comprising:
a memory for storing data and a computer program;
a processor coupled to the memory for executing the computer program, the program comprising instructions for:
receiving a mask of an image of a scene captured by a first one of said cameras;
determining a set of pixels in the mask likely to be in a foreground of an image captured by a second one of said cameras based on the received mask, calibration information, and a geometry of the scene;
generating a set of background pixels for the second camera based on the determined set of pixels;
transmitting the generated set of background pixels to the second camera; and
segmenting the image of the scene captured by the second camera using the transmitted background pixels.
10. An apparatus for segmenting an image of a scene captured using one of a plurality of cameras in a network, the apparatus comprising:
means for receiving a mask of an image of a scene captured by a first one of said cameras;
means for determining a set of pixels in the mask likely to be in a foreground of an image captured by a second one of said cameras based on the received mask, calibration information, and a geometry of the scene;
means for generating a set of background pixels for the second camera based on the determined set of pixels;
means for transmitting the generated set of background pixels to the second camera; and
means for segmenting the image of the scene captured by the second camera using the transmitted background pixels.
11. A non-transitory computer readable medium for segmenting an image of a scene captured using one of a plurality of cameras in a network, the program comprising:
code for receiving a mask of an image of a scene captured by a first one of said cameras;
code for determining a set of pixels in the mask likely to be in a foreground of an image captured by a second one of said cameras based on the received mask, calibration information, and a geometry of the scene;
code for generating a set of background pixels for the second camera based on the determined set of pixels;
code for transmitting the generated set of background pixels to the second camera; and
code for segmenting the image of the scene captured by the second camera using the transmitted background pixels.