Some content of this application is unavailable at the moment.
If this situation persist, please contact us atFeedback&Contact
1. (WO2018069826) A METHOD OF VIEW FRUSTUM DETECTION, CORRESPONDING SYSTEM AND COMPUTER PROGRAM PRODUCT
Note: Text based on automatic Optical Character Recognition processes. Please use the PDF version for legal matters

CLAIMS

1. A method including receiving (100) from at least one image sensor (W) an image signal indicative of the head pose of at least one person, the method including processing said image signal by:

- estimating (102), as a function of said image signal received (100) from at least one image sensor (W) , said head pose of said at least one person,

- generating (104) from said estimated head pose a view frustum of said at least one person,

- generating (106) from said view frustum a visual frustum of attention or VFOA of said at least one person, and

- generating from said visual frustum of attention at least one of:

- a predicted movement trajectory of said at least one person, or

- an estimated attention map for said at least one person as a function of said visual frustum of attention.

2. The method of claim 1, including tracking said at least one person as a function of said predicted movement trajectory.

3 . The method of claim 1 or claim 2, including identifying in said attention map at least one object

(L) to be controlled (C) .

4. The method of claim 3, including controlling by a control device (C) said at least one object (L) identified in said attention map.

5. The method of claim 3 or claim 4, including identifying in said attention map at least one lighting device (L) to be controlled.

6. The method of claim 5, including controlling said at least one lighting device (L) by a control device (C) , said control device (C) preferably including one of:

- a switch,

- a mobile communication device,

- a human gesture sensor.

7. The method of claim 5 or claim 6, including:

- providing a plurality of lighting devices (L) controllable by one control device (C) ,

- identifying in said attention map at least one lighting device (L) to be controlled in said plurality of lighting devices, and

- controlling via said one control device (C) the lighting device identified in said plurality of lighting devices in said attention map.

8. The method of any of the previous claims, including, for a plurality of persons in a scene:

generating (202) a respective personal view frustum for each person in said plurality of persons,

- computing (204) an average group view frustum of the respective personal frustums of a group of persons in said plurality of persons,

- generating (206) an interaction frustum for one person in said group by subtracting from said group frustum the view frustums related to persons in said group which:

- i) are within the highest region of the view frustum of said one person,

- ii) are pointing their view frustum towards said one person, and

iii) are closer than a distance threshold to said one person, and

- cumulating over time the interaction frustum for said one person thereby producing an attention map for said one person.

9. The method of claim 8, including calculating an estimated destination point for one person as a peak of said attention map.

10. The method of claim 9, including:

- detecting (102) the head orientation of said one person, and

- calculating said estimated destination point as a peak in said attention map best matching said head orientation detected (102) .

11. The method of claim 9 or claim 10, including tracking said one person by:

- providing an attention map and an estimated destination point for said one person,

and

- predicting a next location for said one person as a function, preferably as a linear function, of the current location, the head pose, the current velocity and a predicted velocity of said one person.

12. A system including:

- at least one image sensor (W) for generating an image signal indicative of the head pose of at least one person,

- a processing system (1000) coupled with said at least one image sensor (W) for receiving from said at least one image sensor (W) said image signal, the processing system (1000) configured for processing said image signal according to the method of any of claims 1 to 11.

13. A computer program product, loadable in the memory of at least one processing system (1000) and including software code portion for performing the method of any of claims 1 to 11 when the product is run on at least one processing system.