Scene preset identification using quadtree decomposition analysis
Techniques are disclosed for matching a current background scene of an image received by a surveillance system with a gallery of scene presets that each represent a previously captured background scene. A quadtree decomposition analysis is used to improve the robustness of the matching operation when the scene lighting changes (including portions containing over-saturation/under-saturation) or a portion of the content changes. The current background scene is processed to generate a quadtree decomposition including a plurality of window portions. Each of the window portions is processed to generate a plurality of phase spectra. The phase spectra are then projected onto a corresponding plurality of scene preset image matrices of one or more scene preset. When a match between the current background scene and one of the scene presets is identified, the matched scene preset is updated. Otherwise a new scene preset is created based on the current background scene.
1. A computer-implemented method, comprising:
receiving a background scene;
generating a quadtree decomposition of the background scene;
determining whether a window portion of the quadtree decomposition is invalid;
discarding the window portion if the window portion is determined to be invalid;
determining whether the background scene matches a previously captured background scene, based on the quadtree decomposition;
updating the previously captured background scene when the background scene matches the previously captured background scene; and
creating a new background scene when the background scene does not match the previously captured background scene.
2. The computer-implemented method of claim 1 , further comprising:
computing a phase spectrum for each window portion of the quadtree decomposition;
projecting the phase spectra onto corresponding window portions of a quadtree decomposition of the previously captured background scene; and
computing a reconstruction score for the background scene that indicates how closely the background scene matches the previously captured background scene.
3. The computer-implemented method of claim 2 , wherein the computing of the reconstruction score comprises:
computing a reconstruction score for each window portion of the background scene, to produce a plurality of reconstruction scores;
selecting a subset of reconstruction scores from the plurality of reconstruction scores; and
computing the reconstruction score as a weighted sum of the reconstruction scores from the subset of reconstruction scores.
4. The computer-implemented method of claim 2 , wherein the updating of the previously captured background scene comprises:
replacing phase spectra of an oldest background scene that was used to compute the previously captured background scene with the phase spectra for each window portion of the background scene; and
recomputing a quadtree decomposition of the previously captured background scene to produce the new background scene that incorporates the background scene.
5. The computer-implemented method of claim 1 , wherein the background scene is determined to match the previously captured background scene based on a minimum error threshold.
6. The computer-implemented method of claim 1 , wherein determining whether the window portion of the quadtree decomposition is invalid includes:
determining that the window portion is invalid if the window portion is over-saturated.
7. The computer-implemented method of claim 1 , wherein determining whether the window portion of the quadtree decomposition is invalid includes:
determining that the window portion is invalid if the window portion is under-saturated.
8. The computer-implemented method of claim 1 , wherein the creating of the new background scene comprises:
computing a phase spectrum for each window portion of the quadtree decomposition in a sequence of background scenes that includes the background scene; and
performing an analysis to produce image matrices that represent the background scene and the additional background scenes in the sequence.
9. The computer-implemented method of claim 1 , wherein each one of the plurality of window portions of the quadtree decomposition has an equal pixel resolution.
10. A non-transitory computer-readable storage medium containing a program which, when executed by a processor, performs an operation comprising:
receiving a background scene;
generating a quadtree decomposition of the background scene;
determining whether a window portion of the quadtree decomposition is invalid;
discarding the window portion if the window portion is determined to be invalid;
determining whether the background scene matches a previously captured background scene, based on the quadtree decomposition;
updating the previously captured background scene when the background scene matches the previously captured background scene; and
creating a new background scene when the background scene does not match the previously captured background scene.
11. The non-transitory computer-readable storage medium of claim 10 , further comprising:
computing a phase spectrum for each window portion of the quadtree decomposition;
projecting the phase spectra onto corresponding window portions of a quadtree decomposition of the previously captured background scene; and
computing a reconstruction score for the background scene that indicates how closely the background scene matches the previously captured background scene.
12. The non-transitory computer-readable storage medium of claim 11 , wherein updating the previously captured background scene comprises:
replacing phase spectra of an oldest background scene that was used to compute the previously captured background scene with the phase spectra for each window portion of the quadtree decomposition of the background scene; and
recomputing a quadtree decomposition of the previously captured background scene to produce the new background scene that incorporates the background scene.
13. The non-transitory computer-readable storage medium of claim 10 , wherein determining whether the window portion of the quadtree decomposition is invalid includes:
determining that the window portion is invalid if the window portion is over-saturated.
14. The non-transitory computer-readable storage medium of claim 10 , wherein the creating of the new background scene comprises:
computing a phase spectrum for each window portion of the quadtree decomposition in a sequence of background scenes that includes the background scene; and
performing an analysis to produce image matrices that represent the background scene and the additional background scenes in the sequence.
15. A system, comprising:
a video input source configured to capture images;
a processor; and
a memory containing a program, which, when executed on the processor is configured to perform an operation for identifying a scene preset for a background scene of an image captured by the video input source, the operation comprising:
receiving a background scene;
generating a quadtree decomposition of the background scene;
determining whether a window portion of the quadtree decomposition is invalid;
discarding the window portion if the window portion is determined to be invalid;
determining whether the background scene matches a previously captured background scene, based on the quadtree decomposition;
updating the previously captured background scene when the background scene matches the previously captured background scene; and
creating a new background scene when the background scene does not match the previously captured background scene.
16. The system of claim 15 , wherein the operation further comprises:
computing a phase spectrum for each window portion of quadtree decomposition;
projecting the phase spectra onto corresponding window portions of a quadtree decomposition of the previously captured background scene; and
computing a reconstruction score for the background scene that indicates how closely the background scene matches the previously captured background scene.
17. The system of claim 16 , wherein the computing of the reconstruction score comprises:
computing a reconstruction score for each window portion of the background scene, to produce a plurality of reconstruction scores;
selecting a subset of reconstruction scores from the plurality of reconstruction scores; and
computing the reconstruction score as a weighted sum of the best reconstruction scores.
18. The system of claim 16 , wherein the updating of the previously captured background scene comprises:
replacing phase spectra of an oldest background scene that was used to compute the previously captured background scene with the phase spectra for each window portion of the quadtree decomposition of the background scene; and
recomputing a quadtree decomposition of the previously captured background scene to produce the new background scene that incorporates the background scene.
19. The system of claim 15 , wherein the background scene is determined to match the previously captured background scene based on a minimum error threshold.
20. The system of claim 15 , wherein the creating of the new background scene comprises:
computing a phase spectrum for each window portion of the quadtree decomposition in a sequence of background scenes that includes the background scene; and
performing an analysis to produce image matrices that represent the background scene and the additional background scenes in the sequence.