Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N5/00—Details of television systems
  • H04N5/222—Studio circuitry; Studio devices; Studio equipment
  • H04N5/262—Studio circuits, e.g. for mixing, switching-over, change of character of image, other special effects ; Cameras specially adapted for the electronic generation of special effects
  • H04N5/272—Means for inserting a foreground image in a background image, i.e. inlay, outlay
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T7/00—Image analysis
  • G06T7/10—Segmentation; Edge detection
  • G06T7/11—Region-based segmentation
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
  • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
  • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
  • G06V40/18—Eye characteristics, e.g. of the iris
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
  • H04N7/00—Television systems
  • H04N7/14—Systems for two-way working
  • H04N7/141—Systems for two-way working between two video terminals, e.g. videophone
  • H04N7/142—Constructional details of the terminal equipment, e.g. arrangements of the camera and the display
  • H04N7/144—Constructional details of the terminal equipment, e.g. arrangements of the camera and the display camera and display on the same optical axis, e.g. optically multiplexing the camera and display for eye to eye contact
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T2207/00—Indexing scheme for image analysis or image enhancement
  • G06T2207/30—Subject of image; Context of image processing
  • G06T2207/30196—Human being; Person
  • G06T2207/30201—Face

Definitions

• a correction of the viewing direction can in principle be achieved using various methods (B. Kellner et.al., "Video telephone with eye contact?”, NTZ Vol. 38 (1985) No. 10).
• a partially transparent mirror is attached in front of the monitor in such a way that the viewer can see the monitor image through the mirror but with reduced brightness.
• the camera can record the viewer from the direction of the monitor via this mirror, so that recording and playback is possible via an axis (convergence axis).
• the display of the partner picture is taken over by a video projector on a projection screen.
• the camera is located behind the projection screen and can record the viewer through a window provided with a light valve.
• the partner image is reproduced via a video projector on a partially transparent projection screen with a reflective grid structure.
• the camera can be viewed through the permeable structure of the projection screen through the participant take the viewpoint accurate; the contrast range of recording and playback is reduced.
• the display of the partner picture is taken over by a monitor as with a conventional videophone.
• the viewer is recorded on the edge of the monitor from several angles using several cameras.
• a video processor is used to calculate an image from the individual partial images that enables eye contact (J. Liu and R. Skerjanc, "Videophone with a virtual camera for making eye contact", 12th DAGM Symposium, Oberkochen-Aalen, 1990).
• the invention has for its object a method for To indicate correction of the viewing direction on the videophone, which avoids the problems of the methods known from the literature. This object is achieved with the aid of a method for correcting the viewing direction of people on digital images with features according to claim 1.
• the pupils and eyeballs of an imaged person are segmented in the image plane.
• the segments corresponding to the pupils are shifted in the image plane, and areas that become free are filled with the color of the eyeballs.
• a synthetic image of the person depicted is generated, which, however, in contrast to the method (D) known from the literature, is limited to the eyes of the person depicted and is based on considerably simplified models. This method is therefore considerably simpler and less expensive to implement than the methods known from the literature.
• the application of the invention is not limited to the area of the video telephone. Rather, the method can be applied to all digital images and image sequences in which the viewing direction of a person depicted is important. The method is therefore also used in television, e.g. to correct the viewing direction of news speakers and speakers.
• FIG. 1 serves to clarify the problem on which the invention is based.
• Figure 2 shows the geometric relationships in the triangle monitor, camera and eye.
• Figure 3 shows an image section with a human eye and an associated horizontal gray value curve.
• FIG. 4 shows the geometric relationships in the optical illustration.
• FIG. 5 shows an image on which a person with the viewing direction to be corrected is visible.
• FIG. 6 shows an image on which a person is shown with a corrected viewing direction.
• FIG. 1 shows the geometric relationships when a person is photographed with the aid of a camera which is arranged on the side next to a monitor on which the conversation partner of the person being photographed is shown. Since both interlocutors do not look into the camera during the conversation, but look at the monitor, there is an incorrect viewing angle ⁇ between the viewing direction and the camera axis. As can be seen from FIG. 1, eye contact between the conversation partners is not possible with this arrangement.
• FIG. 5 shows a person who was recorded under the conditions visible in FIG. 1. The person depicted does not look at the viewer of the picture, but looks sideways next to the viewer. In the following, a method is described with which the viewing direction of the person depicted is corrected.
• the position of the eyes of a person depicted in the image plane is generally completely unknown. Therefore, the position of the pupils and eyeballs of a person depicted in the image plane must be determined at the beginning of the method. Since the position of the pupils and eyeballs changes over time, it is generally useful and associated with considerable advantages in terms of computational effort if the position of the pupils and eyeballs is tracked over time using suitable methods. In this way it is possible to predict the position of the pupils and eyeballs in the image plane for any given time from the known positions at earlier times. With the help of such predicted estimated values for the locations, it is possible to determine the location at any time with considerably less effort.
• the image areas belonging to the pupils and eyeballs are segmented in the image plane.
• the segments corresponding to the pupils are now shifted in the image plane. Thereby, areas in the area of the eyeballs are cleared which are filled with the color of the eyeballs.
• the position of the pupils and eyeballs can be determined without difficulty.
• various routinely applicable methods in the literature for tracking the movement of the positions of the pupils and eyeballs include the publication by JFSYau and NDDuffy, "Feature Tracking Method For Motion Parameter Estimation In A Model-Based Application", Third International Conference on Image Processing and its Applications, Warwick, UK; July 18-20, 1989; Pp. 531-535.
• JFSYau and NDDuffy "Feature Tracking Method For Motion Parameter Estimation In A Model-Based Application"
• the temporal changes in the position parameters of the pupils and eyeballs can be tracked.
• these methods make it possible to predict estimated values of these position parameters.
• the pupils and eyeballs must be segmented in the image plane.
• the edges of the segments corresponding to the pupils and eyeballs must be determined. These can be found e.g. by determining the gray value curve along a straight line in the image plane which crosses the eye.
• the pupils themselves correspond to the central minimum of such a gray value curve.
• the edges of the pupils thus result with the help of the turning points of this gray value curve, which directly adjoin the central minimum.
• the edges of the eyeball correspond to the minima in the gray value course, which are directly adjacent to the central minimum corresponding to the pupil on both sides.
• the turning points and secondary minima of the gray value curve required to determine the pupil edges or the eyeball edges can be determined in a simple manner with a known gray value curve.
• the gray value curve can be determined along several straight lines crossing the eyes. In this way, the exact segmentation of the pupils and the
• FIG. 3 shows a typical gray value curve along a horizontal straight line through humans liche eye, which is shown in the image above the gray value curve.
• the pupil corresponds to a central minimum in the gray value curve
• the pupil edges correspond to the turning points in the gray value curve directly adjacent to this central minimum.
• the edges of the eyeball correspond to the adjacentlumina of the gray value curve directly adjacent to the central minimum on both sides.
• Figures 2 and 4 show the geometric relationships in the triangle monitor, camera, eye center or in the optical image.
• the distance d between the eye center and the monitor to the eye radius L behaves like the distance p between the camera and monitor to the real pupil shift a.
• the real pupil shift corresponds to the difference in position of the pupils when changing the viewing direction from the monitor to the camera.
• d denotes the distance between the monitor and the center of the eyeball
• p the distance between the monitor and camera
• L the radius of the eyeball
• f the focal length of the camera lens
• r the resolution of the camera, measured in pixels per unit length.
• a low-pass filter can be applied to the changed digital image in a post-processing step in order to eliminate any interfering edges that may occur after the displacement. In this way, disruptive edges, i.e. Eliminate discontinuity contours in the gray value run.

Landscapes

• Engineering & Computer Science (AREA)
• Multimedia (AREA)
• General Physics & Mathematics (AREA)
• Physics & Mathematics (AREA)
• Theoretical Computer Science (AREA)
• Signal Processing (AREA)
• Ophthalmology & Optometry (AREA)
• Human Computer Interaction (AREA)
• Health & Medical Sciences (AREA)
• General Health & Medical Sciences (AREA)
• Computer Vision & Pattern Recognition (AREA)
• Image Processing (AREA)
• Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)

Priority Applications (4)

Applications Claiming Priority (2)

Publications (1)

Family

ID=6424102

Family Applications (1)

Country Status (5)

Cited By (2)

Families Citing this family (24)

Citations (3)

Family Cites Families (12)

• 1991
  • 1991-01-31 DE DE4102895A patent/DE4102895C1/de not_active Expired - Fee Related
• 1992
  • 1992-01-07 US US08/094,171 patent/US5499303A/en not_active Expired - Fee Related
  • 1992-01-07 EP EP92901943A patent/EP0569381B1/de not_active Expired - Lifetime
  • 1992-01-07 JP JP4501637A patent/JPH06504654A/ja active Pending
  • 1992-01-07 WO PCT/DE1992/000007 patent/WO1992014340A1/de not_active Ceased
  • 1992-01-07 DE DE59206573T patent/DE59206573D1/de not_active Expired - Fee Related

Patent Citations (3)

Non-Patent Citations (3)

Also Published As

Similar Documents

Legal Events