BACKGROUND OF INVENTION Viewer measurement plays an important role for statistical surveys of market product survey, museum artwork, advertisement and other similar needs. Firstly, market surveys provide information relating to viewer viewing direction. Secondly, market surveys allow understanding of preferable target object In most typical viewer measurement systems, only one target object is monitored by installing a camera on the target object and analyzing viewer pattern of potential viewers. US 4858000 describes a viewer measurement system using image detection of human features to obtain viewer's view pattern and determine whether a viewer is observing the target object. Unfortunately, is restricted to one viewer at a time which can be time consuming.

US 20090217315 describes a system uses attempts to cater problem by using additional overhead camera installed above target object location and points downwards to detect viewers in vicinity of the target object. However, only one object can be viewed at each time. Another limitation of typical viewer measurement systems is occlusion may happen in a crowd of viewers. In this case, some faces or heads of viewers may not be taken into + account in the viewer measurement systems which affect the effectiveness of the systems.

Therefore, there is a need for a solution for viewer measurement system that solves the problems of occlusion where multiple viewers can be measured simultaneously.

SUMMARY OF INVENTION

Accordingly, there is provided a method of detecting viewer attention to target objects in a + scene, the method includes the steps of identifying each target object position as input image, determining potential region for each viewed target object, validating a moving object as a viewer, determining viewed target object of a validated viewer and obtaining total count for each target object by accumulation.

The present invention consists of several novel features and a combination of parts hereinafter fully described and illustrated in the accompanying description and drawings, it being understood that various changes in the details may be made without departing from the scope of the invention or sacrificing any of the advantages of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will be fully understood from the detailed description given herein + below and the accompanying drawings which are given by way of illustration only, and thus are not limitative of the present invention, wherein:

Figure 1 shows a flowchart depicting components of a method of detecting viewer attention to target objects in a scene in the preferred embodiment of the invention;

Figure 2 shows a flowchart in the Target Object Region Computation component showing the step of determining potential region for each viewed target object in the preferred embodiment of the invention;

Figure 3 shows a graphical representation of the step of determining potential region for each viewed target object in the Target Object Region Computation component in the preferred embodiment of the invention;

Figure 4 shows a flowchart of validating a moving object as a viewer in the Viewer Validation component in the preferred embodiment of the invention;

Figure 5 shows a flowchart showing the step of determining the viewed target object of the valid viewer which is in a Viewed Target Object Determination component in the preferred embodiment of the invention;

Figure 6 shows a graphical representation of Target object positions which are manually provided to identify a region of potential viewed particular target object;

Figure 7 shows a graphical representation of viewed target object referring to a corresponding potential viewed region for each target object; and Figure 8 shows a flow chart showing the step of obtaining total count for each target object in the preferred embodiment of the invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

The present invention relates to a method of detecting viewer attention to target objects in a scene. Hereinafter, this specification will describe the present invention according to the preferred embodiment of the present invention. However, it is to be understood that limiting the description to the preferred embodiment of the invention is merely to facilitate discussion of the present invention and it is envisioned that those skilled in the art may devise various modifications and equivalents without departing from the scope of the appended claims.

The following detailed description of the preferred embodiment will now be described in accordance with the attached drawings, either individually or in combination.

Figure 1 shows a flowchart depicting components of a method of detecting viewer attention to target objects in a scene. The method includes the steps of identifying each target object position as input image, determining potential region for each viewed target object, validating a moving object as a viewer, determining viewed target object of a validated viewer and obtaining total count for each target object by accumulation. Further, the step of determining potential region for each viewed target object further includes the steps of retrieving boundary of each target object, retrieving adjacent boundary of each target object and generating a map of region of interest for each target object Target object positions are manually provided to identify a region of potential viewed particular target object as seen in Figure 6. This step is performed to partition region of interest into segments that represents a possibility of a viewer to be viewing any of the target objects based on spatial information. These segments or partitions assume mat if a viewer exists at a particular region, mere is a possibility of viewing that particular target objects. Auto calculation to determine the potential region is prepared by a Target + Object Region Computation component in this embodiment

The nest step of validating a moving object as a viewer is done by validating the moving object whether the moving object should be considered as a viewer by using a Viewer Validation Component in this embodiment as seen in Figure 1. Subsequently, regions are determined from Target Object Region Computation component and are used in Viewed Target Object Determination component in order to determine the viewed target object of the valid viewer. Finally, the viewed target object from each valid viewer is accumulated frame by frame to obtain a total count for each target object. Analysis is completed by an Information Analysis component in this embodiment of the invention.

Figure 2 shows a flowchart in the Target Object Region Computation component showing the step of determining potential region for each viewed target object Before performing the step, each target object position in the scene is identified manually as an input image for the Target Object Region Computation component. First, boundaries of each target object are retrieved and adjacent boundary is obtained. This is done by computing x coordinate if all the target objects are distributed parallel to x axis and computing y coordinate if all the target objects are distributed parallel to y axis between two target objects in order to draw an equal distance division between the two target objects. Finally, a map of potentially viewed region for each target object is generated. Figure 3 illustrates this step as a graphical representation.

+ Figure 4 shows a flowchart of validating a moving object as a viewer in the Viewer Validation component From a plurality of image sequences, the moving objects are detected and tracked by motion tracking means whereby each object is assigned an identifier. Satisfaction of a plurality of conditions are required, wherein the conditions include the viewer must be within a region of interest and moving objects must be static or with slight movement for a predetermined duration. Hie region of interest includes any potential viewed regions for any target object. As for the moving objects, if the moving objects fulfill both conditions, the moving object is considered a valid viewer and is analyzed in Viewed Target Object Determination component in this embodiment.

Figure 5 shows a flowchart showing the step of determining the viewed target object of the valid viewer which is in a Viewed Target Object Determination component in this embodiment. An occupied region of each valid viewer is determined by referring to a potential viewed region for each target object map as seen in Figure 6 which was generated previously. The target object map includes a plurality of regions which are divided into near and far regions. Near and far regions are differentiated by an empirical threshold. Then, each near and far regions are partitioned into regions accordingly to correspond to target object regions. For example, if the valid viewer is found in a near region, the viewed target object refers to a corresponding potential viewed region for each target object as shown in Figure 7. If the valid viewer exists in a far region, the valid viewer's average moving direction from previous frames are calculated, men the average moving direction is validated by potential viewed region for each target Tf the direction does not relate to any relevant potential viewed target object, then the viewed target object of a particular valid object is considered by a corresponding target object region.

+ Using identification of viewed target object of each valid viewer, a count of each target object is increased accordingly. The Information Analysis component is used to identify a total count for each target object in die scene. For example, a most viewed target object is the target object with a maximum count Figure 8 illustrates a flow chart showmg the step of obtaining total count for each target object The method is used for viewer attention detection for target objects using a single camera. The target objects may be, but not restricted to, paintings in a museum, billboards in a shopping complex and products in a supermarket In order to resolve problems which is occlusion caused by crowd and single monitored target the method uses a top view camera to monitor the scene.

This invention is adapted for use for in estimating number of viewers for each target object as well as provide additional information such as attention span of each viewer at any target object. The disclosed invention is suitable, but not restricted to, for use in detecting viewer attention.