代码拉取完成，页面将自动刷新
2019-05-01 v6.0.0：
[TOC]
This document shows you how to integrate the Faceunity Nama SDK into your Android project.
This section describes the structure of the Demo file, the various directories, and the functions of important files.
+FULiveDemoDroid
+app // app
+src
+main
+assets // Assets
+image // source image
+effects // props files
+bundle // bundle files
+res // icon
+java // Java source code
+makeup // makeup resource
+material // stickers
+res // icon
+poster // Facefusion
+res // App resource files
+faceunity // faceunity module
+libs // nama.jar
+src
+main
+assets // Nama resource files
+java // Java source code
+jniLibs // Nama so
+docs // contents
+README.md // README
Android API 18 or above，GLES 2.0 or above
Android Studio 3.0 or above，GLES 2.0 or above
The certificate issued by the Android platform is the authpack.java file. If you have obtained the authentication certificate, put the certificate file in the project faceunity module com.faceunity.fulivedemo package.
Initialize SDK environment, load SDK data and perform authentication. Initialization must be performed before other interface, otherwise it will lead to an exception.
public static native int fuSetup(byte[] v3data, byte[] authdata);
Parameter description:
v3data : v3.bundle file directory
authdata : secret key array, must be set properly so SDK can work normally
Example code：
InputStream v3 = context.getAssets().open(BUNDLE_v3);
byte[] v3Data = new byte[v3.available()];
v3.read(v3Data);
v3.close();
faceunity.fuSetup(v3Data, authpack.A());
Notes：only needs to be setup once after startup, where the authpack.A() key array is declared in authpack.java.
Prop item creation interface:：
public static native int fuCreateItemFromPackage(byte[] data);
Parameter description:：
data :prop item binary data
Return value：
int : prop handle
Sometimes in practical application multiple prop items will be utilized at the same time, in that case parameter that Nama SDK's video processing interfaces accept is a int array including multiple prop item handles, so we need to creat a int array to store these prop item handles. Next we will create a face beautification prop item handle and store it in position ITEM_ARRAYS_EFFECT in int array. Example code as below:
InputStream is = mContext.getAssets().open(bundle.path());
byte[] itemData = new byte[is.available()];
int len = is.read(itemData);
is.close();
mItemsArray[ITEM_ARRAYS_EFFECT] = faceunity.fuCreateItemFromPackage(itemData);
updateEffectItemParams( mItemsArray[ITEM_ARRAYS_EFFECT]);
Single Prop Item Destruction:
public static native void fuDestroyItem(int item);
Parameter description：
item :prop item handle to be destroyed
This interface will release resource corresponding to imported handle. Example code as below ：
if (mItemsArray[ITEM_ARRAYS_EFFECT] > 0)
faceunity.fuDestroyItem(mItemsArray[ITEM_ARRAYS_EFFECT]);
All Prop Items Destruction
public static native void fuDestroyAllItems();
This interface can destroy resources corresponding to all prop item handle, please set this handle to 0 after processing as well. Example code as below:
Arrays.fill(mItemsArray, 0);
faceunity.fuDestroyAllItems();
If certain position of handle array needs to be switched, new prop item handle needs to be created first, and then switch this handle with the one needs to be replaced, at last destroy replaced handle. Below is an example of replacing the position ITEM_ARRAYS_EFFECT handle of handle array:
final Effect effect = (Effect) msg.obj;
final int newEffectItem = loadItem(effect);
queueEvent(new Runnable() {
@Override
public void run() {
if (mItemsArray[ITEM_ARRAYS_EFFECT] > 0) {
faceunity.fuDestroyItem(mItemsArray[ITEM_ARRAYS_EFFECT]);
}
mItemsArray[ITEM_ARRAYS_EFFECT] = newEffectItem;
setMaxFaces(effect.maxFace());
}
});
due to the asynchronous loading, using the queueEvent method to update the handle. (The queueEvent mechanism is the same as the queueEvent mechanism of GLSurfaceView)
Notes: If you destroy the old props first, and then create new props, you may have a stuttering phenomenon.
Send video image data and prop item handle together to Nama SDK's rendering interfaces and prop items' effect will be rendered to video image after rendering process.
public static native int fuDualInputToTexture(byte[] img, int tex_in, int flags, int w, int h, int frame_id, int[] h);
Parameter description:
img :video image data byte[], support NV21(default), I420SP, RGBA format
tex_in video image texture ID
flags flags, may specify img data format, return texture ID's prop item mirroring and so on
w :video image data's width
h :video image data's height
frame_id: current processing video frame's serial number
items :int array including multiple prop item handle
Return value:
int :processed video image data texture ID
Example code as below:
int flags = mInputTextureType | mInputImageFormat;
if (mCurrentCameraType != Camera.CameraInfo.CAMERA_FACING_FRONT)
flags |= FU_ADM_FLAG_FLIP_X;
int fuTex = faceunity.fuDualInputToTexture(img, tex, flags, w, h, mFrameId++, mItemsArray);
public static native int fuRenderToNV21Image(byte[] img, int w, int h, int frame_id, int[] items, int flags);
Parameter description:
img :video image data byte[], processed image data will be written back to this byte[]
w :video image data's width
h :video image data's height
frame_id :current processing video frame's serial number
items :int array including multiple prop item handle
flags flags, may specify img data format, return texture ID's prop item mirroring and so on
Return value:
int :processed video image data texture ID
Example code as below：
int flags = mInputImageFormat;
if (mCurrentCameraType != Camera.CameraInfo.CAMERA_FACING_FRONT)
flags |= FU_ADM_FLAG_FLIP_X;
int fuTex = faceunity.fuRenderToNV21Image(img, w, h, mFrameId++, mItemsArray, flags);
Beautification
Configuration of face beautification function is similar to that of adding special effect prop items to video image, create face beautification prop handle first and then save it in handle array.
InputStream beauty = mContext.getAssets().open(BUNDLE_face_beautification);
byte[] beautyData = new byte[beauty.available()];
beauty.read(beautyData);
beauty.close();
mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX] = faceunity.fuCreateItemFromPackage(beautyData);
When processing video image, beautification prop item handle will be imported to video image rendering interface through handle array, face beautification effect will be added to original video image after rendering.
Example code as below：
int fuTex = faceunity.fuDualInputToTexture(img, tex, flags, w, h, mFrameId++, mItemsArray);
Parameter Configuration Face beautification prop item mainly contain seven modules: filter, skin whitening, ruddy, blur, eye brighten, tooth whiten and face outline beautification. Each module has its default effect, parameters which can be adjusted are as belows:
The filter function is mainly controlled by the parameters filter_level and filter_name.
filter_name : refer to the documentation：Beautification Filters User Specification ，detailed description of the corresponding key value of the filter.
filter_level : value range 0.0-1.0, 0.0 is no effect, 1.0 is maximum effect, default value is 1.0, sample code：
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "filter_name", mFilterName.filterName());
filter_name : the value is a string, the default value is "origin", which is the use of the original image effect, sample code：
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "filter_level", mFaceBeautyFilterLevel);
Control skin whitening level through parameter color_level
The color_level value ranges from 0.0 to 1.0, 0.0 is no effect, 1.0 is the maximum effect, and the default value is 0.2.
Parameter setting example as below:
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "color_level", mFaceBeautyColorLevel);
Control ruddy level through parameter red_level
Red_level ranges from 0.0-1.0, 0.0 is no effect, 1.0 is maximum effect, and the default value is 0.5.
Parameter setting example as below:
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "red_level", mFaceBeautyRedLevel);
There are 4 parameters to control skin blur：blur_level，skin_detect，nonshin_blur_scale，heavy_blur。
blur_level: specifies blur level，This parameter's recommended value range is 0~6，the default value is 6.0skin_detect: specifies whether enable skin detection，0 is on ，1 is offnonskin_blur_scale: specifies non skin area blur weaken level when enable skin detection，This parameter's range is 0~1，the default value is 0.45heavy_blur: specifies whether enable heavy blur function，0 is light blur，1 is heavy blur.Parameter setting example code as below:
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "skin_detect", mFaceBeautyALLBlurLevel);
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "heavy_blur", mFaceBeautyType);
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "blur_level", 6 * mFaceBeautyBlurLevel);
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "nonshin_blur_scale", 0.45);
Eye brighten level can be controlled through parameter eye_bright
eye_bright :this parameter's recommended value range is 0~1，1.0 is the maximum effect, the default value is 1.0.
Parameter setting example code as below:
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "eye_bright", mBrightEyesLevel);
Tooth Whiten level can be controlled through parameter tooth_whiten .
tooth_whiten: This parameter's recommended value range is 0~1，1.0 is the maximum effect, the default value is 1.0.
Parameter setting example code as below:
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "tooth_whiten", mBeautyTeethLevel);
Face outline beautification supports four basic beautification types: Goddess, cyber celebrity, nature, default and one advanced face beautification type: custom, which specified by parameter face_shape: Goddess(0),cyber celebrity(1), nature(2), default(3) and custom(4).
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "face_shape", mFaceBeautyFaceShape);
Based on above mentioned four basic face outline beautification and one advanced face outline beautification, we provide following three parameters: face_shape_level, eye_enlarging, cheek_thinning in addition.
face_shape_level controls levels changing to specified basic face shape. This parameter's recommended value range is 0~1 which 0 stands for no effect, disable face outline beautification，1.0 is the maximum effect, the default value is 1.0.
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "face_shape_level", mFaceShapeLevel);
The parameter eye_enlarging is used to control the size of the eye. This parameter is affected by the parameter face_shape_level. The recommended value range for this parameter is [0, 1]. Greater than 1 is the enhancement effect, the default is 0.5.
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "eye_enlarging", mFaceBeautyEnlargeEye);
The parameter cheek_thinning is used to control the size of the face. This parameter is affected by the parameter face_shape_level. The recommended value range for this parameter is [0, 1]. Greater than 1 is the enhancement effect, the default is 0.5.
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_FACE_BEAUTY_INDEX], "cheek_thinning", mFaceBeautyCheekThin);
face outlines adjustment
Added optimized face-lifting and big-eye effects, increased forehead adjustment, chin adjustment, thin nose, and mouth shape adjustment. 4 face_shape is set to 4 to enable fine face adjustment. FULiveDemo can be selected from the face. Define to turn on fine face adjustment.
Instructions：
face_shape: 4, // 4 is enable advance function，0～3 is basic functionFace-lifting
Optimize the effect, more natural than before
Instructions：
face_shape: 4, // 4 is enable advance function，0～3 is basic function
cheek_thinning: 0.0, // Use the original parameter cheek_thinning to control face-lifting, range 0 - 1, default 0.0Eye Brighten
Optimize eye brighten effect , more natural than before
Instructions：
face_shape: 4, // 4 is enable advance function，0～3 is basic function
eye_enlarging: 0.0, // Use the original parameter eye_enlarging to control large eyes, range 0 - 1, default 0.5Forehead adjustment
A new beauty prop that can adjust the forehead size
Instructions：
face_shape: 4, // 4 is enable advance function，0～3 is basic function
intensity_forehead: 0.5, // Greater than 0.5 becomes larger, less than 0.5 becomes smaller, default is 0.5Chin adjustment
A new beauty prop that can adjust the chin size
Instructions：
face_shape: 4, // 4 is enable advance function，0～3 is basic function
intensity_chin: 0.5, // Greater than 0.5 becomes larger, less than 0.5 becomes smaller, default is 0.5Thin nose
A new beauty prop that can adjust the thin size
Instructions：
face_shape: 4, //4 is enable advance function，0～3 is basic function
intensity_nose: 0.0, // 0 is the normal size, greater than 0 starts thin nose, range 0 - 1, default 0.0Mouth adjustment
A new beauty prop that can adjust the mouth size
Instructions：
face_shape: 4, // 4 is enable advance function，0～3 is basic function
intensity_mouth: 0.5, // Greater than 0.5 becomes larger, less than 0.5 becomes smaller, default is 0.5Gesture recognition function is realized by loading prop item as well currently. Gesture recognition prop items cotain gesture to be recognized, animation effect triggered when gesture recognized and control script. Loading process is same with that of common prop items and face beautification prop item.
In LiveDemo heart_v2.bundle is love heart gesture demo prop item. Load it as prop item to render shall enable gesture recognition function. Gesture recognition prop item can exist with common and face beautification prop item at the same time, just like expand mItemsArray to three in face beautification and load gesture recognition prop item at last.
Process of customize gesture prop is same with that of 2D prop item, please contact Faceunity technical support for details.
Note: Some of the new version of the gesture props need to use the non-lite version of the SDK in order to use normally.
High efficient full screen anti-aliasing, makes 3D render effect smoother.
Instructions：
InputStream animoji3D = mContext.getAssets().open(BUNDLE_animoji_3d);
byte[] animoji3DData = new byte[animoji3D.available()];
animoji3D.read(animoji3DData);
animoji3D.close();
mItemsArray[ITEM_ARRAYS_EFFECT_ABIMOJI_3D] = faceunity.fuCreateItemFromPackage(animoji3DData);
Performe accurate face rebuild through photograph, then it will support real time face expression driving and preset face expression play. It can be used in real time application or be used to creat face expression package.
This function contains two methods to creat resources:
Utilize FUEditor v4.3.0 and above to creat off line props
Utilize cloud service provided by Faceunity to upload photograph to creat props in real time.
Please contact our technical support to get detailed information about cloud service.
Instructions:
added 5 exaggeration transformation(Face Warp Function) prop items.
Instructions：
Music filter is powered by time-stamp of playing music, it can be utilized by introducing music playing course to music filter prop item every time processing video image, example code as below:
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_EFFECT_INDEX], "music_time", time);
You can simulate music playing course if there is no music. Prop item in demo corresponds to music of 28s(28000ms) in length, time counting can be started from loading music filter prop item, get time difference between current time and time when starting loading music filter every time before processing video image and convert it into ms, set time to 0 when time difference is over 28000ms. You can check FULiveDemo for effect and create one utilizing FUEditor(v4.2.1 and above).
The Nama SDK supports the Avatar pinch feature from 6.0.0.
The virtual character supports the pinch function. By using the bundle generated by the FUEditor 6.0.0 template facepup, the corresponding parameters can be adjusted on the client side, so that the user can customize the face, hairstyle, lip color and other parts, generate and save the avatar desired by the user. The avatar can be driven by the engine in real time according to the face, and the expression is synchronized with the face.
After loading the prepared pinch props, you can set the following parameters to pinch the face.
Enter the pinch state
//Enter the pinch state
faceunity.fuItemSetParamd(mItemsArray[ITEM_ARRAYS_EFFECT_INDEX],"enter_facepup",1);
Clear all pinch parameters
//Clear all pinch parameters
faceunity.fuItemSetParamd(mItemsArray[ITEM_ARRAYS_EFFECT_INDEX],"clear_facepup",1);
Set the pinch parameter
//Set the weight of the first pinch attribute "big eye", the range [0-1]. param corresponds to the first pinch face attribute here, starting from 1.
faceunity.fuItemSetParamd(mItemsArray[ITEM_ARRAYS_EFFECT_INDEX],"\{\"name\":\"facepup\",\"param\":\"1\"\}",1);
//Set the weight of the second pinch attribute "tip chin", range [0-1]. param corresponds to the first pinch face attribute here, starting from 1
faceunity.fuItemSetParamd(mItemsArray[ITEM_ARRAYS_EFFECT_INDEX],"\{\"name\":\"facepup\",\"param\":\"2\"\}",0.5);
**Save or Exit **
Choose one of the two
//1. Directly exit the pinch state, do not save the current pinch state, and enter the tracking state. Use the last pinch face to track facial expressions.
faceunity.fuItemSetParamd(mItemsArray[ITEM_ARRAYS_EFFECT_INDEX],"quit_facepup",1);
//2. Trigger to save the pinch face, and exit the pinch state, enter the tracking state. Time-consuming operation, set if necessary.
fuItemSetParamd(mItemsArray[ITEM_ARRAYS_EFFECT_INDEX],"need_recompute_facepup",1);
Comic Filter
Comic Filter is filter in comic style, utilizing method is same with common prop item, just load and save it in prop item handle array and then introduce to video image processing interface. In our demo, we utilize comic filter with Animoji in AR mode, but it's not limited in this kind of utilization, you can also utilize it alone or with any other prop item together. Please pay attention that Nama SDK will set comic filter's 'glVer' property according to current OpenGL ES version. Only utilizing OpenGL ES 3.0 can get good comic filter effect, below versions may cause bad effect. Setting method code as below:
faceunity.fuItemSetParam(itemHandle, "glVer", 3); // 2 stands for compatibility mode, 3 stands for normal mode。
Animoji in AR Mode
Unlike Animoji in normal mode, AR mode will show reality scenes and Animoji will move following face to add more interactivities. Utilizing Animoji in AR mode together with comic filtetr can make Animoji and reality scene's style unitive and blend very natural. Animoji in AR mode setting method code as below:
// Set Animoji to follow the face
faceunity.fuItemSetParam(itemHandle, "{\"thing\":\"<global>\",\"param\":\"follow\"}", 1); // A value of 1 means open, and a value of 0 means closed.
Nama SDK started supporting tongue driving function from v5.6. When utilizing prop item with tongue effect, loading tongue driving data model first is necessary and loading method code is as below:
InputStream tongue = context.getAssets().open(BUNDLE_tongue);
byte[] tongueDate = new byte[tongue.available()];
tongue.read(tongueDate);
tongue.close();
faceunity.fuLoadTongueModel(tongueDate);
Nama SDK started supporting face fusion function from v5.8. This function can perform fusion effect of user's face to model's face in the poster to realize face transfer effect.
First load change_face.bundle, and then please refer to Face Transfer Interface Documentation，also you can refer to our demo.
Nama SDK started supporting new version make up function from v5.8. This function can add lip, blush, eye brow, eye shadow, eye liner, eyelash and eye pupil effects and support exquisite modification.
First load face_makeup.bundle, and then please refer toMake Up Parameter Specification，also you can refer to our demo.
The Nama SDK supports the hairdressing function from 5.8.0. This function can change the hair color of the characters. Currently, it supports 8 kinds of common hair color and 5 kinds of gradient colors, and adjusts the hairdressing degree.
First load hair_color.bundle, then use the parameter Index to switch the color. The recommended value range of this parameter is 07. The intensity of the hair color can be adjusted by the parameter Strength. The recommended value range of this parameter is 01. Example code:
/**Set hairdressing parameters**/
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_EFFECT_HAIR_NORMAL_INDEX], "Index", mHairColorIndex);
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_EFFECT_HAIR_NORMAL_INDEX], "Strength", mHairColorStrength);
First load hair_gradient.bundle, then use the parameter Index`` to switch the color. The recommended value range of this parameter is 04. The intensity of the hair color can be adjusted by the parameter Strength. The recommended value range of this parameter is 01. Example code:
/**Set hairdressing parameters**/
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_EFFECT_HAIR_GRADIENT_INDEX], "Index", mHairColorIndex);
faceunity.fuItemSetParam(mItemsArray[ITEM_ARRAYS_EFFECT_HAIR_GRADIENT_INDEX], "Strength", mHairColorStrength);
The Nama SDK supports texture and beauty features starting with 6.0.0.
The Light makeup is a more sophisticated and efficient beauty solution, which includes four modules: blur, beauty, filter and beauty. It provides 60+ sets of rich material library to support customers to switch styles and effects.
First load light_makeup.bundle, then set parameters such as blush, eye shadow, eyeliner, lipstick, etc.，and you can refer to our sample code in FULiveDemo.
-keep class com.faceunity.wrapper.faceunity {*;}
此处可能存在不合适展示的内容，页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容，可点击提交进行申诉，我们将尽快为您处理。