Face Recognition Algorithm¶

REVISION HISTORY¶

Revision No.	Description	Date
100	Initial release	04/27/2023
101	Add Face Attribute Recognition	07/04/2023
102	Sync with latest api & add error code	05/20/2024

1. Algorithm Description¶

Face recognition mainly distinguishes faces and determines whether the captured face is from a whitelisted person. The entire algorithm includes face detection, face attribute recognition, face expression recognition, facial key points, face filtering, face tracking, face alignment, feature extraction, and face comparison.

Among them, face attributes can determine five attributes of the face: glasses, gender, mask, mustache, and age; face expressions can identify seven expressions, which are normal, happy, sad, surprised, scared, disgusted, and angry.

Algorithm Accuracy

ALGO	Threshold	43	44	45	46	47	48	49	50	51	52	53	54	55	56	57	58
FR	TAR%	99.4	98.9	98.5	98.1	97.6	96.9	96.4	95.2	93.9	93.4	92.1	90.9	89.1	86.7	85.4	84.2
FR	FAR%	2.3	1.8	1	0.7	0.4	0.1	0	0	0	0	0	0	0	0	0	0

Model Introduction

model	Function	Resolution(w*h)	Input Format
fr_det_y24s.img	Face Detection	480*288	yuvsp420_nv12
fr_det_y36s.img	Face Detection	640*352	yuvsp420_nv12
fr_det_y48s.img	Face Detection	800*480	yuvsp420_nv12
fr_feature_as.img	Face Feature Extraction (Small)	112*112	BGRA8888
fr_feature_am.img	Face Feature Extraction (Medium)	112*112	BGRA8888
fr_feature_all.img	Face Feature Extraction (Large)	112*112	BGRA8888
fr_angle_y66.img	Face Angle Estimation	64*64	yuvsp420_nv12
fr_cos256.img	Face Feature Comparison Model	256*512	INT16
far101_224y_emo.img	Face Emotion Recognition Model	224*224	yuvsp420_nv12
far101_224y_emo_s45.img	Face Emotion Recognition Model (Small)	224*224	yuvsp420_nv12
far103_224y.img	Face Attribute Recognition Model	224*224	yuvsp420_nv12

2. Interface Calling Process¶

Face recognition mainly involves two processes: one is the face registration process, which creates the face whitelist (database). The other is the face recognition process, which identifies the captured face to determine if it is from a whitelisted person.

2.1. Registration Process¶

The interface call for the registration process: ALGO_FR_Init → ALGO_FR_CreateHandle → ALGO_FR_Detect → ALGO_FR_Align → ALGO_FR_FeatureExtract. Finally, the extracted features are saved to become the database.

2.2. Recognition Process¶

The interface call for the recognition process is: ALGO_FR_Init → ALGO_FR_CreateHandle → ALGO_FR_Detect → ALGO_FR_Align → ALGO_FR_FeatureExtract → ALGO_FR_FeatureCompare, comparing the features from the captured image with the features in the database to determine whether they belong to the same person.

2.3. Face Attribute Recognition Process¶

The interface call for face attribute recognition is: ALGO_FR_Init → ALGO_FR_CreateHandle → ALGO_FR_Detect → ALGO_FR_Attr, using the captured face image for face attribute recognition.

Note: The face image sent to this interface after cropping needs to be expanded by a certain range, the expansion width and height are: 1.25*bbox_w, 1.25*bbox_h

2.4. Multi-Detector Recognition Process¶

First Detector:

ALGO_FR_Init (Only needs to be done once) --> ALGO_FR_CreateHandle (Create handle using the detection model path specified in InitFrParam_t) --> ALGO_FR_Detect --> ALGO_FR_Align --> ALGO_FR_FeatureExtract --> ALGO_FR_FeatureCompare

Subsequent Detectors:

ALGO_FR_AddHandle (Create a new handle by specifying the path to the new detection model) --> ALGO_FR_Detect --> ALGO_FR_Align --> ALGO_FR_FeatureExtract --> ALGO_FR_FeatureCompare

Note: The largest detector should be used as the first detector!

3. Function Module API¶

API Name	Function
ALGO_FR_Init	SDK Initialization
ALGO_FR_CreateHandle	Create Handle
ALGO_FR_Detect	Face Detection
ALGO_FR_Attr	Face Attribute Recognition
ALGO_FR_Align	Face Alignment
ALGO_FR_FeatureExtract	Feature Extraction
ALGO_FR_FeatureCompare	Feature Comparison
ALGO_FR_ReleaseHandle	Release Handle
ALGO_FR_Cleanup	Release Entire SDK Environment
ALGO_FR_HandleResetParam	Reset Algorithm Configurable Parameters
ALGO_FR_FaceQuanlity	Perform Face Angle Estimation and Quality Filtering
ALGO_FR_BatchFeatureCompare	Perform Batch Face Feature Comparison (requires a copy of the comparison features)
ALGO_FR_BatchFeatureCompareV2	Perform Batch Face Feature Comparison (no need to copy comparison features)
ALGO_FR_AddHandle	Create Handle for New Detection Model

3.1. ALGO_FR_Init¶

Function

Initialize the SDK, only needs to be initialized once.

Syntax

MI_S32 ALGO_FR_Init(InitFrParam_t initParam);

Parameters

Parameter Name Description Input/Output

initParam Initialization Parameters Input
Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

Example

{
    int ret=0;
    InitFrParam_t initParam;
    memcpy(initParam.ipu_firware_bin,"/config/dla/ipu_firmware.bin", IPU_MAX_LENGTH);
    memcpy(initParam.det_model_path, "./models/fr_det_y24s.img", MODEL_MAX_LENGTH);
    memcpy(initParam.attr_model_path, "./models/far101_224y_s45.img", MODEL_MAX_LENGTH);
    memcpy(initParam.emo_model_path, "./models/far101_224y_emo_s45.img", MODEL_MAX_LENGTH);
    memcpy(initParam.feature_model_path, "./models/fr_feature_as.img", MODEL_MAX_LENGTH);
    memcpy(initParam.cos_model_path, "./models/i6c/fr_cos256.img", MODEL_MAX_LENGTH);
    initParam.box_min_size = 5;
    initParam.det_thredhold = 0.5;
    initParam.attr_thredhold = 0.3;
    initParam.eye_distance = 20;
    initParam.filter_angle_ratio = 0.2;
    initParam.fr_mode = FR_MODE_FACE_DET | FR_MODE_FACE_DET_RECOG | FR_MODE_FACE_ATTR | FR_MODE_FACE_EMOTION;
    ret= ALGO_FR_Init(initParam);
}

3.2. ALGO_FR_CreateHandle¶

Function

Create a handle; one thread creates one handle.

Syntax

MI_S32 ALGO_FR_CreateHandle(MI_S64* detectorId)

Parameters

Parameter Name Description Input/Output

detectorId Handle Pointer Input
Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.3. ALGO_FR_Detect¶

Function

Face Detection.

Syntax

MI_S32 ALGO_FR_Detect(MI_S64 detectId, AlgoFrInputInfo_t *stBufInfo, MI_S32 width, MI_S32 height, ParamDet_t* params, DetectBox_t** detectout, MI_S32* facecount)

Parameters

Parameter Name	Description	Input/Output
detectorId	Handle	Input
stBufInfo	Data Pointer (yuv420_nv12) (Size 480x288)	Input
width	Width of the input raw image	Input
height	Height of the input raw image	Input
params	Configuration parameters for detection	Input
detectOut	Output for detected face box	Output
faceCount	Number of detected face boxes	Output

Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.4. ALGO_FR_Attr¶

Function

Face Attribute Recognition.

Syntax

MI_S32 ALGO_FR_Attr(MI_S64 detectId, AlgoFrInputInfo_t *stBufInfo, FaceAttr_t *results);

Parameters

Parameter Name Description Input/Output

detectorId Handle Input

stBufInfo Data Pointer (argb8888) (Size 224x224) Input

results Face Attribute Results Output
Note

The face buffer sent to this interface needs to be expanded by a certain range, the expansion width and height are: +0.25*bbox_w, +0.25*bbox_h
Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.5. ALGO_FR_Align¶

Function

Face Alignment.

Syntax

MI_S32 ALGO_FR_Align(MI_U8* imageData, MI_S32 width, MI_S32 height, MI_S32 type, DetectBox_t detectOut, MI_U8* outData);

Parameters

Parameter Name	Description	Input/Output
imageData	Original Image Data Pointer	Input
width	Width of the original image	Input
height	Height of the original image	Input
type	Image type, 1 for yuv420_nv12, 0 for argb8888	Input
detectOut	Face Box	Input
outData	Output aligned 112x112 ARGB888 Data, cannot be empty	Output

Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.6. ALGO_FR_FeatureExtract¶

Function

Face Feature Extraction.

Syntax

MI_S32 ALGO_FR_FeatureExtract(MI_S64 detectId, MI_U8* imageData, MI_S16* featureOut);

Parameters

Parameter Name	Description	Input/Output
detectId	Handle
imageData	Input image data output from Align interface (112x112)	Input
featureOut	Face's feature data, 512-length MI_S16 feature	Output

Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.7. ALGO_FR_FeatureCompare¶

Function

Feature Comparison.

Syntax

MI_S32 ALGO_FR_FeatureCompare(MI_S16* feature1, MI_S16* feature2, MI_S32 length, MI_FLOAT* similarity);

Parameters

Parameter Name	Description	Input/Output
feature1	Features of the first face	Input
feature2	Features of the second face	Input
length	Feature length, currently 512	Input
similarity	Similarity between the two faces	Output

Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.8. ALGO_FR_ReleaseHandle¶

Function

Release Handle.

Syntax

MI_S32 ALGO_FR_ReleaseHandle(MI_S64 detectorId);

Parameters

Parameter Name Description Input/Output

detectorId Handle Input
Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.9 ALGO_FR_Cleanup¶

Function

Release the entire SDK environment.
Syntax
```
MI_S32 ALGO_FR_Cleanup();
```
Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.10 ALGO_FR_HandleResetParam¶

Function

Reset the algorithm's configurable parameters

Syntax

MI_S32 ALGO_FR_HandleResetParam(MI_S64 detectorID, AlgoFrFilterParam_t param);

Parameters

Parameter Name Description Input/Output

detectorID Handle Input

param Algorithm parameters, see AlgoFrFilterParam_t Input
Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.11 ALGO_FR_FaceQuanlity¶

Function

Perform face angle estimation and quality filtering.

Syntax

MI_S32 ALGO_FR_FaceQuanlity(MI_S64 detectId, AlgoFrInputInfo_t* stBufInfo, DetectBox_t* box, AlgoFaceQuanlityParam_t param, FaceQuanlityAngle_t* angle);

Parameters

Parameter Name	Description	Input/Output
detectorID	Handle	Input
stBufInfo	Frame data pointer (yuv420_nv12) (Size 64x64)	Input
box	Pointer for output face detection box	Input/Output
param	Face quality parameters, see AlgoFaceQuanlityParam_t	Input
angle	Output result of face angle, see FaceQuanlityAngle_t	Output

Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.12 ALGO_FR_BatchFeatureCompare¶

Function

Perform batch face feature comparison (requires a copy of the comparison features)

Syntax

MI_S32 ALGO_FR_BatchFeatureCompare(MI_S64 detectId, MI_S16* feature1, MI_S16* feature2, MI_S32 batch, MI_FLOAT* similarity);

Parameters

Parameter Name	Description	Input/Output
detectorID	Handle	Input
feature1	Target feature for comparison (1)	Input
feature2	Source features for comparison (batch)	Input
batch	Number of source features for comparison	Input
similarity	Similarity output of feature comparison	Output

Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.13 ALGO_FR_BatchFeatureCompareV2¶

Function

Perform batch face feature comparison (no need to copy comparison features)

Syntax

MI_S32 ALGO_FR_BatchFeatureCompareV2(MI_S64 detectId, AlgoFrInputInfo_t* feature1, AlgoFrInputInfo_t* batchFeature, MI_S32 batch, MI_FLOAT* similarity);

Parameters

Parameter Name	Description	Input/Output
detectorID	Handle	Input
feature1	Target feature for comparison (1), passed in MMA buffer format	Input
feature2	Source features for comparison (batch), passed in MMA buffer format	Input
batch	Number of source features for comparison	Input
similarity	Similarity output of feature comparison	Output

Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

3.14. ALGO_FR_AddHandle¶

Function

Add a handle for the new detector, allowing specification of the detection model path so that each handle can use a different face detector.

Syntax

MI_S32 ALGO_FR_AddHandle(MI_S64* detectorID, const MI_U8* det_model_path);

Parameters

Parameter Name Description Input/Output

detectorId Handle Pointer Input

det_model_path Path to detection model Input
Return Value

Return Value Description

0 Success

Others Failure (See Error Code)
Dependencies

Header File: algo_fr_api.h

Library File: libsstaralgo_fr.so, libsstaralgo_fr.a

4. Data Types¶

Data Type	Definition
ALGO_FrMode_e	Algorithm Mode.
InitFrParam_t	Initialization Parameters for Face Algorithm.
ParamDet_t	Algorithm Detection Input Parameters for Selecting Face Detection Mode.
AlgoFrInputInfo_t	Algorithm Input Data Structure.
FrPoint_t	Defines Key Point Coordinates.
DetectBox_t	Detection Box Information.
FaceAttr_t	Face Attribute Category Information.
AlgoFrFilterParam_t	Face Filtering Parameter Structure
FaceQuanlityAngle_t	Face Angle Output Structure
AlgoFaceQuanlityParam_t	Face Quality Parameter Structure

4.1. ALGO_FrMode_e¶

Description

Algorithm Mode.

Definition

typedef enum
{
    FR_MODE_FACE_DET,
    FR_MODE_FACE_DET_RECOG,
    FR_MODE_FACE_DET_ATTR,
    FR_MODE_FACE_EMOTION,
} ALGO_FrMode_e;

Members

Member Name Description

FR_MODE_FACE_DET Face Detection

FR_MODE_FACE_RECOG Face Recognition

FR_MODE_FACE_ATTR Face Attribute Recognition

FR_MODE_FACE_EMOTION Face Emotion Recognition
Related Data Types and Interfaces

ALGO_FR_Init

4.2. InitFrParam_t¶

Description

Initialization Parameters for Face Algorithm.

Definition

typedef struct InitFrParam_
{
    MI_U8 ipu_firware_bin[IPU_MAX_LENGTH];
    MI_U8 det_model_path[MODEL_MAX_LENGTH];
    MI_U8 feature_model_path[MODEL_MAX_LENGTH];
    MI_U8 cos_model_path[MODEL_MAX_LENGTH];
    MI_U8 attr_model_path[MODEL_MAX_LENGTH];
    MI_U8 emo_model_path[MODEL_MAX_LENGTH];
    MI_FLOAT det_thredhold;
    MI_FLOAT attr_thredhold;
    MI_S32   box_min_size;
    MI_FLOAT filter_angle_ratio;
    MI_FLOAT eye_distance;
    MI_U8   fr_mode;
    MI_BOOL had_create_device;
} InitFrParam_t;

Members

Member Name	Description
ipu_firware_bin	Path of the Firmware_bin
det_model_path	Path of the Detection Model
feature_model_path	Path of the Feature Extraction Model
cos_model_path	Path of the Similarity Model
attr_model_path	Path of the Face Attribute Recognition Model
emo_model_path	Path of the Face Emotion Recognition Model
det_thredhold	Threshold for detection, recommended to set to 0.5
attr_thredhold	Threshold for face attribute recognition, recommended to set to 0.5
box_min_size	Minimum detection box size, recommended to set to 10
filter_angle_ratio	Ratio for filtering face angle, recommended to set to 0.23
eye_distance	Distance between eyes, recommended to set to 30
fr_mode	Algorithm Mode, defaults to face recognition mode
had_create_device	Whether a device has been created for multiple handles

Related Interfaces

ALGO_FR_Init

4.3. ParamDet_t¶

Description

Algorithm Detection Input Parameters for Selecting Face Detection Mode.

Definition

typedef struct ParamDet_
{
    MI_S32 datatype;
}ParamDet_t;

Members

Member Name	Description
datatype	0 indicates detect the maximum face; 1 indicates detect and take all faces; 2 indicates video input, take the maximum face and add tracking id; 3 indicates video input, take all faces and add tracking id

Related Data Types and Interfaces

ALGO_FR_Detect

4.4. AlgoFrInputInfo_t¶

Description

Algorithm Input Data Structure.

Definition

typedef struct AlgoFrInputInfo_
{
    void* pt_tensor_data;
    MI_PHY phy_tensor_addr;
    MI_U32 bufsize;
    MI_S64 pts;
} AlgoFrInputInfo_t;

Members

Member Name Description

pt_tensor_data Virtual address of input data

phy_tensor_addr Physical address of input data

bufsize Size of input data

pts Timestamp of input data
Related Data Types and Interfaces

ALGO_FR_Detect

4.5. FrPoint_t¶

Description

Defines Key Point Coordinates.

Definition

typedef struct FrPoint_
{
    MI_FLOAT x;
    MI_FLOAT y;
}FrPoint_t;

Members

Member Name Description

x x-coordinate of the key point

y y-coordinate of the key point
Related Data Types and Interfaces

DetectBox_t

4.6. DetectBox_t¶

Description

Detection box information.

Definition

typedef struct DetectBox_
    {
        MI_FLOAT x1;
        MI_FLOAT y1;
        MI_FLOAT x2;
        MI_FLOAT y2;
        MI_FLOAT score;
        FrPoint_t landmark[FR_POINT_LEN];
        MI_S32 face_id;
        MI_S32 befiltered;
    }DetectBox_t;

Members

Member Name	Description
x1	x-coordinate of the top-left point of the face box
y1	y-coordinate of the top-left point of the face box
x2	x-coordinate of the bottom-right point of the face box
y2	y-coordinate of the bottom-right point of the face box
score	Confidence of the face
landmark	Five key points of the face
face_id	Tracking id of the face
befiltered	Whether the face is filtered, True if filtered

Related Data Types and Interfaces

ALGO_FR_Detect

4.7. FaceAttr_t¶

Description

Attribute category information.

Definition

typedef struct FaceAttr_t_
{
    MI_BOOL   eye_glass;
    MI_BOOL   male;
    MI_BOOL   mask;
    MI_S32    age;
    MI_BOOL   mustache;
    MI_S32    emotion;
}FaceAttr_t;

Members

Member Name Description

eye_glass Glasses

male Male

mask Mask

age Age

mustache Mustache

emotion Emotion
Related Data Types and Interfaces

ALGO_FR_Attr

4.8. AlgoFrFilterParam_t¶

Description

Face filtering parameter structure.

Definition

typedef struct AlgoFrFilterParam_
{
    MI_S32   box_min_size;
    MI_FLOAT filter_angle_ratio;
    MI_FLOAT eye_distance;
    MI_FLOAT det_thredhold;
    MI_FLOAT attr_thredhold;
    MI_FLOAT motion_sensitive;   //[0.0-1.0]
}AlgoFrFilterParam_t;

Members

Member Name	Description
box_min_size	Minimum face detection box size, default 20 (pixels)
filter_angle_ratio	Head tilt/pitch angle requirement (0.0~1.0), larger values require more straight angles, default is 0.25
eye_distance	Minimum interpupillary distance, default is 20 (pixels)
det_thredhold	Minimum face detection score, default is 0.3
attr_thredhold	Score threshold for face attribute recognition
motion_sensitive	Required motion magnitude of the face (0.0~1.0), larger values require less motion of the face

Related Data Types and Interfaces

ALGO_FR_HandleResetParam

ALGO_FR_FaceQuanlity

4.9. FaceQuanlityAngle_t¶

Description

Face angle output structure.

Definition

typedef struct FaceQuanlityAngle_
{
    MI_FLOAT pitch;
    MI_FLOAT yaw;
    MI_FLOAT roll;
    MI_FLOAT quanlity;
}FaceQuanlityAngle_t;

Members

Member Name Description

pitch Face pitch angle

yaw Face yaw angle

roll Face roll angle

quanlity Comprehensive score of face angle (value range 0.0~1.0)
Related Data Types and Interfaces

ALGO_FR_FaceQuanlity

4.10. AlgoFaceQuanlityParam_t¶

Description

Face quality parameter structure.

Definition

typedef struct AlgoFaceQuanlityParam_
{
    MI_FLOAT thredhold;
}AlgoFaceQuanlityParam_t;

Members

Member Name Description

thredhold Threshold for face quality score (value range 0.0~1.0)
Related Data Types and Interfaces

ALGO_FR_FaceQuanlity

5. Error Codes ¶

Error Code	Value	Description
E_ALGO_SUCCESS	0	Operation Successful
E_ALGO_HANDLE_NULL	1	Algorithm Handle is Null
E_ALGO_INVALID_PARAM	2	Invalid Input Parameter
E_ALGO_DEVICE_FAULT	3	Hardware Error
E_ALGO_LOADMODEL_FAIL	4	Model Loading Failed
E_ALGO_INIT_FAIL	5	Algorithm Initialization Failed
E_ALGO_NOT_INIT	6	Algorithm Not Initialized
E_ALGO_INPUT_DATA_NULL	7	Algorithm Input Data is Null
E_ALGO_INVALID_INPUT_SIZE	8	Invalid Algorithm Input Data Dimension
E_ALGO_INVALID_LICENSE	9	Invalid License
E_ALGO_MEMORY_OUT	10	Memory Insufficient
E_ALGO_FILEIO_ERROR	11	File Read/Write Operation Error
E_ALGO_INVALID_OUTPUT_SIZE	12	Invalid Algorithm Output Data Dimension

Parameter Name	Description	Input/Output
detectorId	Handle Pointer	Input
det_model_path	Path to detection model	Input

Member Name	Description
FR_MODE_FACE_DET	Face Detection
FR_MODE_FACE_RECOG	Face Recognition
FR_MODE_FACE_ATTR	Face Attribute Recognition
FR_MODE_FACE_EMOTION	Face Emotion Recognition

Member Name	Description
pt_tensor_data	Virtual address of input data
phy_tensor_addr	Physical address of input data
bufsize	Size of input data
pts	Timestamp of input data

Member Name	Description
x	x-coordinate of the key point
y	y-coordinate of the key point

Member Name	Description
eye_glass	Glasses
male	Male
mask	Mask
age	Age
mustache	Mustache
emotion	Emotion

Member Name	Description
pitch	Face pitch angle
yaw	Face yaw angle
roll	Face roll angle
quanlity	Comprehensive score of face angle (value range 0.0~1.0)

Return Value	Description
0	Success
Others	Failure (See Error Code)

Return Value	Description
0	Success
Others	Failure (See Error Code)

Return Value	Description
0	Success
Others	Failure (See Error Code)

Return Value	Description
0	Success
Others	Failure (See Error Code)