please select

Gesture Recognition

Overview

Input the camera's OpenGL texture and output real-time gesture detection data. You can use this data for further development.

Android Integration Guide

Integrate Beauty AR SDK on Android, for details please refer to: Integrating SDK (Android).

Android Interface Invocation

1. Turn on the gesture detection feature switch (in XmagicApi.java)
public void setFeatureEnableDisable(String featureName, boolean enable);
Fill in featureName with XmagicConstant.FeatureName.HAND_DETECT, and set enable to true.
2. Set data callback (in XmagicApi.java)
void setAIDataListener(XmagicApi.OnAIDataListener aiDataListener)

public interface OnAIDataListener {
void onFaceDataUpdated(List<TEFaceData> faceDataList);
void onHandDataUpdated(List<TEHandData> handDataList);
void onBodyDataUpdated(List<TEBodyData> bodyDataList);
void onAIDataUpdated(String data);
}
onAIDataUpdated returns a JSON structured string data.

Callback JSON Data Description

In the callback JSON data, the gesture-related data is in "hand_info", and the format is as follows:
"hand_info": {
"gesture": "PAPER",
"hand_point_2d": [180.71888732910156, 569.2958984375, ... , 353.8714294433594, 836.246826171875]
}
The explanations of each field in hand_info are as follows:
Field
Explanation
gesture
Gesture Type Name
hand_point_2d
Captured gesture data information
The following gestures are currently supported:
Order
Gesture
Type Name
Example Image
1
Heart
HEART

img


2
Gestrue with number 5(open)
PAPER

img


3
Gesture with number 2
SCISSOR

img


4
Fist
FIST

img


5
Gesture with number 1
ONE

img


6
I love you
LOVE

img


7
Thumb up
LIKE

img


8
OK
OK

img


9
Rock
ROCK

img


10
Gesture with number 6
SIX

img


11
Gesture with number 8
EIGHT

img


12
Lift
LIFT

img


13
Gesture with number 3
THREE

img


14
Gesture with number 4
FOUR

img


If it is an undetected gesture, the gesture type name is OTHER.