Intention identification method

• US 6,154,558 A
• Filed: 04/22/1998
• Issued: 11/28/2000
• Est. Priority Date: 04/22/1998
• Status: Expired due to Term

• Current Assignee: Hon Hai Precision Industry Co., Ltd.
• Original Assignee: Kuan-Hong Hsieh
• Inventors: Hsieh, Kuan-Hong
• Primary Examiner(s): Johns, Andrew W.
• Assistant Examiner(s): NAKHJAVAN, SHERVIN K
• Application Number:
  US09/064,050
• Time in Patent Office: 951 Days
• Field of Search: 382/100, 382/103, 382/107, 382/286, 382/291, 345/358, 348/143, 348/154, 348/155, 348/156
• US Class Current: 382/103
• CPC Class Codes: G06F 3/017 Gesture based interaction, e.g. based on a set of recognized hand gestures interaction based on gestures traced on a digitiser G06F3/04883
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16

Abstract

An intention identification method, i.e., a method of tracking movement of an object for the purpose of determining intention in order to operate a computer, includes the steps of processing a pattern of an object captured by a video camera and converting the pattern into a plurality of pixels, analyzing a variation trend of the pixels based on variation of the pixels in any direction relative to a preset geometric center, and comparing a quantity representing the direction and moving amount of the pixels to a characteristic curve, the characteristic curve being formed by a non-activate region set with respect to the center of the pixels, and an activate region also set with respect to the center of the pixels and which is larger than the non-activate region.

Claims

1. An intention identification method for tracking and predicting an movement of a pattern of first object captured by a video camera and received by a computer in order to determine an intention of a user and thereby operate the computer, comprising the steps of:

processing the pattern and converting the pattern into a plurality of pixels;

analyzing a variation trend of the pixels, based on variation of the pixels in any direction relative to a preset geometric center of the pixels in order to obtain a quantity representative of a direction and moving amount of the pixels for comparison with a respective characteristic curve so that the intention of the object may be determined; and

operating the computer according to results of the intention determination,

wherein the characteristic curve is formed by a non-activate region set with respect to center point of the pixels, and an activate region set with respect to the center point of pixels and which is larger than that of the non-activate region,

wherein during formation of the characteristic curve, a further quick action region is formed with respect to the center point of the pixels, which is larger than the activate region, and

wherein the step of obtaining a quantity representative of a direction and moving amount of the pixels comprises the steps of;

(a) capturing an image of a background object and transferring the captured background object image to the computer;

(b) setting pixel data extracted from the background object image as reference data, constructing a reference coordinate system including X, Y, and Z coordinate axes with respect to said center of the pixels of the reference data, and analyzing said trend of said pixels in said reference coordinate system, said pixel data being selected based on lack of movement of said pixel data within a predetermined time period;

(c) using an integral trend judging method to detect movement of the first object relative to said X and Y coordinate axes, using a variation ratio of movements relative to the respective X and Y coordinate axes to determine a movement trend for the first object; and

(d) determining that the object has moved if the movement trend is some direction is over a threshold, and deriving said quantity representative of a direction and moving amount of the pixels.

2. The intention method as recited in claim 1, further comprising the steps of filtering a signal with respect to the variation of the direction and displacement of the pixels through a motion filter, extracting a maximum axial vector for a unit time, and removing noise for preventing an error judgment due to interference from background and micro-variation due to slight motion of the object.

3. The intention method as recited in claim 1, further comprising the step of constructing four quadrants defined by said X coordinate axis and Y coordinate axis, and outputting a Z coordinate signal when numbers of the pixels in at least three of said quadrants have changed in a same way, the variation ratio is a variation ratio based on the total number of pixels that have changed, and the variation ratio is over a threshold.

4. The intention method as recited in claim 1, wherein in step (b), the reference data is set to represent an action of human behavior in interacting to the computer'"'"'s instruction.

5. The intention method as recited in claim 1, wherein the movement of the first object is detected in step (c) according to an approximate variation of the center point.

6. The intention method as recited in claim 1, wherein said first object is a human body part and said intention is an intention to operate an operating pattern menu and cursor icon displayed on a display screen of the computer, which operated by the action of the human body part.

7. The intention method as recited in claim 6, wherein the operating pattern displayed ion the display screen is distributed on a corner area of the display screen.

8. The intention method as recited in claim 1, further including the step of presetting at least one reference behavior mode, and then comparing a track formed from a last output signal of the direction and displacement quantity in a unit time with a track of the reference behavior mode and, if the direction and displacement quantity track and the reference behavior mode track are similar, interpreting the reference behavior mode as the intention of the object.