Method and device for retargeting a 3D content

1. Method for retargeting a 3D content comprising at least a first view and a second view the method comprising:

determining a first saliency map from said first view and a second saliency map from said second view using a processor;

determining a first cropping window from said first saliency map and a second cropping window from said second saliency map;

combining said first and second cropping windows into a final cropping window identified with parameters determined from said first and second cropping windows;

determining a forbidden area map indicating regions where final cropping borders are to be avoided;

converting an aspect region of the final cropping window into a target aspect ratio using the forbidden area map; and

cropping said first view and said second view according to said final cropping window, the cropped first and second views forming the retargeted 3D content.

2. Method according to claim 1 , wherein the method further comprises filtering over time the parameters of the final cropping window.

3. Method according to claim 1 , wherein the method further comprises:

filtering over time the parameters of the final cropping window; determining a forbidden area map indicating regions where final cropping window borders are to be avoided; and

converting the aspect ratio of the final cropping window into a target aspect ratio using the forbidden area map and a depth map.

4. Method according to claim 2 , wherein the step of determining the forbidden area map comprises the steps of:

segmenting one of said at least two views into a segmentation map of regions;

computing, for each region of the segmentation map, a saliency value from the saliency map of said one of said at least two views;

thresholding said saliency values computed for each region to form a forbidden area map.

5. Device for retargeting a 3D content comprising at least a first view and a second view the device comprising:

means for determining a first saliency map from said first view and a second saliency map from said second view;

means for determining a first cropping window from said first saliency map and a second cropping window from said second saliency map;

means for combining said first and second cropping windows into a final cropping window identified with parameters determined from said first and second cropping windows;

means for determining a forbidden area map indicating regions where final cropping window borders are to be avoided;

means for converting the aspect ratio of the final cropping window into a target aspect ratio using the forbidden area map; and

means for cropping said first view and said second view according to said final cropping window, the cropped first and second views forming the retargeted 3D content.

6. Device according to claim 5 , wherein the device further comprises

means for filtering over time the parameters of the final cropping window.

7. Device according to claim 5 , wherein the device further comprises:

means for filtering over time the parameters of the final cropping window;

means for determining a forbidden area map indicating regions where final cropping window borders are to be avoided; and

means for converting the aspect ratio of the final cropping window into a target aspect ratio using the forbidden area map and a depth map.

8. Device according to claim 6 , wherein the means for determining the forbidden area map comprises:

means for segmenting one of said at least two views into a segmentation map of regions;

means for computing, for each region of the segmentation map, a saliency value from the saliency map of said one of said at least two views;

means for thresholding said saliency values computed for each region to form a forbidden area map.

9. A non-transitory program storage device, readable by a computer, tangibly embodying a program of instructions executable by the computer to perform a method comprising the steps of

determining a first saliency map from said first view and a second saliency map from said second view;

determining a first cropping window from said first saliency map and a second cropping window from said second saliency map;

combining said first and second cropping windows into a final cropping window identified with parameters determined from said first and second cropping windows;

determining a forbidden area map indicating regions where final cropping borders are to be avoided;

converting an aspect region of the final cropping window into a target aspect ratio using the forbidden area map; and

cropping said first view and said second view according to said final cropping window, the cropped first and second views forming the retargeted 3D content.

10. A device for retargeting a 3D content including at least a first view and a second view, comprising:

a processor configured to determine a first saliency map from said first view and a second saliency map from said second view, to determine a first cropping window from said first saliency map and a second cropping window from said second saliency map, to combine said first and second cropping windows into a final cropping window identified with parameters determine from said first and second cropping windows, to determine a forbidden area map indicating regions where final cropping window borders are to be avoided, to convert the aspect ratio of the final cropping window into a target aspect ratio using the forbidden area map, and to crop said first view and said second view according to said final cropping window, the cropped first and second views forming the retargeted 3D content.

11. The device according to claim 10 , wherein the processor is further configured to filter over time the parameters of the final cropping window.

12. The device according to claim 11 , wherein the processor is further configured to segment one of said at least two views into a segmentation map of regions, to compute, for each region of the segmentation map, a saliency value from the saliency map of said one of said at least two views, and to threshold said saliency values computed for each region to form a forbidden area map.

13. The device according to claim 10 , wherein the processor is further configured to filter over time the parameters of the final cropping window, to determine a forbidden area map indicating regions where final cropping window borders are to be avoided, and to convert the aspect ratio of the final cropping window into a target aspect ratio using the forbidden area map and a depth map.