Gesture recognition method and gesture recognition apparatus using the same

Patent metadata

Field	Value
Publication number	US-2016140384-A1
Application number	US-201514631822-A
Country	US
Kind code	A1
Filing date	Feb 25, 2015
Priority date	Nov 17, 2014
Publication date	May 19, 2016
Grant date	—

Abstract

A gesture recognition method and an gesture recognition apparatus using the same method are proposed. The method includes the following steps: transmitting an infrared signal, and detecting whether an object appears in a predetermined range based on a first infrared signal retrieved from the infrared signal being reflected; when the object appears in the predetermined range, capturing a plurality of images including the object and a plurality of second infrared signals retrieved from the infrared signal being reflected; and recognizing the images to retrieve a horizontal movement direction of a gesture of the object within the predetermined range, and recognizing the second infrared signals to retrieve a depth movement direction of the gesture within the predetermined range.

First claim

What is claimed is:

1 . A gesture recognition method, suitable for a gesture recognition apparatus having an image capturing unit, and comprising:
transmitting an infrared signal, and detecting whether an object appears in a predetermined range based on a first infrared signal retrieved from the infrared signal being reflected; when the object appears in the predetermined range, capturing a plurality of images including the object and a plurality of second infrared signals retrieved from the infrared signal being reflected; and recognizing the images to retrieve a horizontal movement direction of a gesture of the object within the predetermined range, and recognizing the second infrared signals to retrieve a depth movement direction of the gesture within the predetermined range.

2 . The gesture recognition method of claim 1 , wherein
when the object appears in the predetermined range, identifying a current position of the object according to a signal strength of the first infrared signal; and adjusting an image capturing range facing the object based on the current position to retrieve the images and the second infrared signals.

3 . The gesture recognition method of claim 2 , wherein the object is a human body, and the step of adjusting the image capturing range facing the object based on the current position to retrieve the images and the second infrared signals comprises:
executing a face recognition mechanism for the human body to recognize a face position; and using a reference point in the face position to serve as a data center point of the image capturing range.

4 . The gesture recognition method of claim 1 , wherein the step of recognizing the second infrared signals to retrieve a depth movement direction of the gesture within the predetermined range comprises:
determining whether an absolute value of a signal strength difference between a t th second infrared signal and a (t+1) th second infrared signal among the second infrared signals is greater than a predetermined threshold; when the absolute value is greater than the predetermined threshold, determining whether a signal strength of the (t+1) th second infrared signal lasts for a predetermined time; when the signal strength of the (t+1) th second infrared signal lasts for the predetermined time, determining whether the signal strength of the (t+1) th second infrared signal is greater than a signal strength of the t th second infrared signal;
if yes, recognizing that the gesture is a pushing gesture; and
if no, recognizing that the gesture is a pulling gesture.

5 . The gesture recognition method of claim 1 , wherein the step of recognizing the images to retrieve the horizontal movement direction of the gesture of the object within the predetermined range comprises:
recognizing a moving palm appearing in the images, and defining a plurality of feature points on the moving palm; retrieving a plurality of estimated moving distances of the feature points in the images; finding a plurality of robust feature points from the feature points based on the estimated moving distances; retrieving a plurality of movement parameters of each of the robust feature points in the images; defining a plurality of moving directions of the robust feature points according to the movement parameters of each of the robust feature points; and deciding a gesture moving direction of the gesture according to the moving directions of the robust feature points.

6 . The gesture recognition method of claim 5 , wherein and the step of recognizing the moving palm appearing in the images comprises:
building a background model based on the images; generating a plurality of gay scale images of the images, wherein a k th gray scale image among the gray scale images is corresponding to a k th image among the images; for a (k+1) th gray scale image and the k th gray scale image, subtracting the k th gray scale image from the (k+1) th gray scale image to generate a gray scale difference image; and finding the moving palm from the gray scale difference image.

7 . The gesture recognition method of claim 5 , wherein an i th feature point among the feature points is corresponding to an i th estimated moving distance among the estimated moving distances, and the step of finding the robust feature points from the feature points based on the estimated moving distances comprises:
when the i th estimated moving distance is greater than a predetermined distance threshold, defining that the i th feature point is one of the robust feature points.

8 . The gesture recognition method of claim 5 , wherein the movement parameters of a j th robust feature point among the robust feature points comprises a total horizontal movement distance, a total vertical movement distance, a horizontal moving start point, a horizontal moving destination, a vertical moving start point and a vertical moving destination in the images, and the j th robust feature point is corresponding to a j th moving direction among the moving directions,
wherein the step of defining the moving directions of the robust feature points according to the movement parameters of each of the robust feature points comprises: for the j th robust feature point, determining whether the total horizontal movement distance is greater than the total vertical movement distance; when the total horizontal movement distance is greater than the total vertical movement distance, determining whether the horizontal moving start point is greater than the horizontal moving destination;
if yes, defining that the j th moving direction is a leftward direction;
if no, defining that the j th moving direction is a rightward direction;
when the total horizontal movement distance is not greater than the total vertical movement distance, determining whether the vertical moving start point is greater than the vertical moving destination;
if yes, defining that the j th moving direction is an upward direction; and
if no, defining that the j th moving direction is a downward direction.

9 . The gesture recognition method of claim 8 , wherein the step of deciding the gesture moving direction of the gesture according to the moving directions of the robust feature points comprises:
counting an amount of each of the upward direction, the downward direction, the leftward direction and the rightward direction among the moving directions; and defining the moving direction with a highest amount among the moving directions is the gesture moving direction of the gesture.

10 . The gesture recognition method of claim 6 , further comprising:
when a proportion of the background model occupied by a foreground in the images is greater than a predetermined threshold, stopping recognizing the gesture.

11 . The gesture recognition method of claim 1 , wherein the gesture recognition apparatus comprises an infrared transmitter for transmitting the infrared signal and the image capturing unit for continuously capturing the images and the second infrared signals, and the step of recognizing the images and the second infrared signals is performed only when detecting that the object appears in the predetermined range.

12 . The gesture recognition method of claim 1 , wherein the gesture recognition apparatus comprises an infrared transmitter for transmitting the infrared signal, an infrared receiver for receiving the first infrared signal and the second infrared signals, and the image capturing unit for capturing the images and the second infrared signals, and the image capturing unit is activated to capture the images and the second infrared signals only when detecting that t

Assignees

• Wistron Corp

Inventors

• Yang Jie-Ci
• Lin Hui-Chen
• Shen Tzu-Chiang
• Kao Meng-Chao

Classifications

• G06F3/0304 Primary

  Detection arrangements using opto-electronic means (constructional details of pointing devices not related to the detection arrangement using opto-electronic means G06F3/033; optical digitisers G06F3/042) · CPC title

• G06V40/28

  Recognition of hand or arm movements, e.g. recognition of deaf sign language (static hand signs G06V40/113) · CPC title

• G06V10/143

  Sensing or illuminating at different wavelengths · CPC title

• H04N23/20

  for generating image signals from infrared radiation only · CPC title

• G06T7/60

  Analysis of geometric attributes · CPC title