Image analysis – Applications – Motion or velocity measuring
Reexamination Certificate
1998-05-28
2002-01-01
Au, Amelia M. (Department: 2623)
Image analysis
Applications
Motion or velocity measuring
C382S154000
Reexamination Certificate
active
06335977
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention is related to an action recognizing apparatus for imaging motion of a human and the like, for specifying motion of this imaged object from an acquired image with respect to each of the portions thereof, and for reproducing the motion of the imaged object based on the specified motion, and also related to a recording medium in which a motion recognizing program is recorded.
As the conventional technique in which a gesture of a human is imaged by an image sensor such as a CCD, the imaged gesture is reflected by motion of a person expressed by computer graphics, and then the original gesture is reproduced on a computer screen. One optical motion capture is described in the publication titled “Imidas '97” at pages 846 to 847. In this optical motion capture, light reflection markers are attached to respective portions of the human body as the imaged object. The motion of these portions is imaged by a camera along different directions, the three-dimensional positions of the light reflection markers at respective time instants are detected by processing the acquired images, and then three-dimensional data about the motion of the hands and the feet are calculated based on the three-dimensional positions. The original motion of the human can be reproduced on a human model expressed by the computer graphics employing these three-dimensional data.
The above-described conventional action recognizing system has the following problems. That is, it is very difficult to attach the light reflection markers specifying the respective portions of the human body onto the respective portions of the human body. Motion of an animal, such as a monkey, other than a human cannot be easily reproduced in accordance with this conventional action recognizing system because it is very difficult to attach the light reflection markers to the animal. There is another problem in that the application range of this conventional action recognizing system is limited.
Also, in accordance with this conventional action recognizing system, since the positions of the light reflection markers at respective time instants are detected to analyze their motion, there are situations in which all motions cannot be acquired, for instance, some of the light reflection markers may be hidden. As a result, after the detection data is manually corrected off-line, producing the motion, the corrected detection data is used to reflect the original motion of the human model. As a consequence, there is a further problem that the motion of the human could not be reflected on the human model in real time.
SUMMARY OF THE INVENTION
The present invention has been made to solve the above-described problem, and, therefore, has an object of providing action recognizing apparatus reproducing motion of an imaged object, such as a human, in real time by analyzing motion of an image produced by imaging the motion of the imaged object without employing an auxiliary means, such as a light reflection marker. Also, another object of the present invention is to provide a recording medium in which an action recognizing program is recorded.
An action recognizing apparatus, according to a first arrangement of the present invention, comprises an image input unit for forming an image containing motion of an imaged object at a plurality of imaging angles with a plurality of image sensors to acquire a plurality of two-dimensional images; a two-dimensional motion detecting unit for calculating a direction and magnitude of motion of each of pixels within each of the two-dimensional images formed by the image input unit; a region specifying unit for specifying a region having a magnitude of motion larger than or equal to a predetermined magnitude, the motion magnitude being calculated by the two-dimensional motion detecting unit; a three-dimensional motion detecting unit for positioning the specific region specified by the region specifying unit with respect to the plurality of two-dimensional images, whereby a three-dimensional position of the specific region and three-dimensional motion of the specific region are calculated; a portion predicting unit for predicting that the three-dimensional motion of the specific region calculated by the three-dimensional motion detecting unit corresponds to any of the portions of the imaging object; and an action reproducing unit for moving a model in a simulation manner based the motion predicted by the portion predicting unit with respect to each of the portions, whereby motion of the imaged object is reproduced on the model.
Also, the region specifying unit in the action recognizing apparatus, according to a second arrangement of the present invention, subdivides a two-dimensional image into a plurality of scanning ranges which overlap with each other; scans each of the scanning ranges to select as a subject region a region having a maximum averaged value of motion magnitude of the respective pixels; and furthermore establishes a region where the averaged value is larger than a predetermined value as the specific region.
Also, the action reproducing unit in the action recognizing apparatus, according to a third arrangement of the present invention, reproduces the motion of the imaging object by way of motion of a model expressed by computer graphics.
Also, the action reproducing unit in the action recognizing apparatus, according to a fourth arrangement of the present invention, reproduces the motion of the imaging object by way of motion of an automatic control machine.
Also, the action reproducing unit in the action recognizing apparatus, according to a fifth arrangement of the present invention, includes a second model, different from the model used to reproduce the motion of the imaged object, whereby a mutual effect between the motion of the imaging operation and the motion of the second model is realized in a simulation.
In a recording medium of the present invention, an action recognizing program causes a computer to calculate a direction and magnitude of motion of each of pixels within two-dimensional images formed at a plurality of imaging angles; specify a region having calculated motion magnitude larger than or equal to a predetermined magnitude; position the specified specific region with respect to the plurality of two-dimensional images, whereby a three-dimensional position of the specific region and three-dimensional motion of the specific region are calculated; predict the three-dimensional motion of the calculated specific region corresponding to any of the portions of the imaged object; and move a model in a simulation based on the motion with respect to each of the predicted portions, whereby motion of the imaging object is reproduced on the model.
REFERENCES:
patent: 4739400 (1988-04-01), Veitch
patent: 5214758 (1993-05-01), Ohba et al.
patent: 5227985 (1993-07-01), DeMenthon
patent: 5495576 (1996-02-01), Ritchey
patent: 5552823 (1996-09-01), Kageyama
patent: 5633993 (1997-05-01), Redmann et al.
patent: 5767861 (1998-06-01), Kimura
patent: 5877778 (1999-03-01), Dow et al.
patent: 6154723 (2000-11-01), Cox et al.
patent: WO-9322762 (1993-11-01), None
K. Rohr, “Towards Model-Based Recognition of Human Movements In Image Sequences”, CVGIP: Image Understanding, vol. 59, No. 1, Jan. 1994, pp. 94-115.
Gavrilla et al., “3-D Model-Based Tracking of Humans in Action: a Multi-View Approach”, in Proc. IEEE Computer Vision and Pattern Recognition, San Francisco, 1996.
Torres et al., “A system for human motion matching between synthetic and real images based on a biomechanic graphical model”, ISBN: 0-8186-6435-5, Nov. 1994.*
Sudhaker et al., “Motion estimation from a sequence of stereo images: a direct method”, ISSN: 0018-9472, Jul. 1994.*
Pentland et al., “Space—time gestures”, ISBN 0-8186-3880-X, Jun. 1993.*
Perales et al, “An automatic two-camera human motion perception system based on biomechanic model matching”, ISBN: 0-7803-0233-8, Oct. 1991.*
Young et al., “3-D motion estimation using a sequence of noisy stereo images: models, estimation, and uniqueness
Au Amelia M.
Leydig , Voit & Mayer, Ltd.
Miller Martin
Mitsubishi Denki & Kabushiki Kaisha
LandOfFree
Action recognizing apparatus and recording medium in that... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Action recognizing apparatus and recording medium in that..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Action recognizing apparatus and recording medium in that... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2836830