6. Data Types¶
6.1. CVI_TDL_Core¶
6.1.1. CVI_TDL_SUPPORTED_MODEL_E¶
【Description】
This enum defines all Deep Learning Models in the TDL SDK.
The following table shows each model Id and its corresponding model functionality:
Index |
Model ID |
Description |
---|---|---|
0 |
CVI_TDL_SUPPORTED_MODEL _YOLOV3 |
80 types of object detection |
1 |
CVI_TDL_SUPPORTED_MODEL _YOLOV5 |
80 types of object detection |
2 |
CVI_TDL_SUPPORTED_MODEL _YOLOV6 |
80 types of object detection |
3 |
CVI_TDL_SUPPORTED_MODEL _YOLOV7 |
80 types of object detection |
4 |
CVI_TDL_SUPPORTED_MODEL _YOLOV8_DETECTION |
80 types of object detection |
5 |
CVI_TDL_SUPPORTED_MODEL _YOLOX |
80 types of object detection |
6 |
CVI_TDL_SUPPORTED_MODEL _PPYOLOE |
80 types of object detection |
7 |
CVI_TDL_SUPPORTED_MODEL
|
80 types of object detection |
8 |
CVI_TDL_SUPPORTED_MODEL _YOLOV8_HARDHAT |
Hardhat detection(YOLOv8) |
9 |
CVI_TDL_SUPPORTED_MODEL _HAND_DETECTION |
Hand detection |
10 |
CVI_TDL_SUPPORTED_MODEL _PERSON_PETS_DETECTION |
Humanoid and cats and dogs detection |
11 |
CVI_TDL_SUPPORTED_MODEL _PERSON_VEHICLE_DETECTION |
Humanoid and vehicle detection |
12 |
CVI_TDL_SUPPORTED_MODEL _HAND_FACE_PERSON_DETECTION |
Hand, face and human -like detection |
13 |
CVI_TDL_SUPPORTED_MODEL _HEAD_PERSON_DETECTION |
Person, face, head target detection |
14 |
CVI_TDL_SUPPORTED_MODEL _MOBILEDETV2_COCO80 |
80 types of object detection |
15 |
CVI_TDL_SUPPORTED_MODEL _MOBILEDETV2_VEHICLE |
Vehicle detection |
16 |
CVI_TDL_SUPPORTED_MODEL _MOBILEDETV2_PEDESTRIAN |
Pedestrian detection |
17 |
CVI_TDL_SUPPORTED_MODEL _RETINAFACE |
Human face detection(RetinaFace) |
18 |
CVI_TDL_SUPPORTED_MODEL _SCRFDFACE |
Human face detection(ScrFD Face) |
19 |
CVI_TDL_SUPPORTED_MODEL _RETINAFACE_IR |
Infrared face detection(RetinaFace) |
20 |
CVI_TDL_SUPPORTED_MODEL _FACEMASKDETECTION |
Face detection of wearing a mask |
21 |
CVI_TDL_SUPPORTED_MODEL _FACELANDMARKERDET2 |
Human face key point detection |
22 |
CVI_TDL_SUPPORTED_MODEL _FACEQUALITY |
The quality of face |
23 |
CVI_TDL_SUPPORTED_MODEL _MASKCLASSIFICATION |
Face mask detection |
24 |
CVI_TDL_SUPPORTED_MODEL _FACERECOGNITION |
Hand detection |
25 |
CVI_TDL_SUPPORTED_MODEL _FACEATTRIBUTE_CLS |
Face attribute |
26 |
CVI_TDL_SUPPORTED_MODEL _HANDCLASSIFICATION |
Gesture recognition |
27 |
CVI_TDL_SUPPORTED_MODEL _HAND_KEYPOINT |
Hand keypoint detection |
28 |
CVI_TDL_SUPPORTED_MODEL _HAND_KEYPOINT_CLASSIFICATION |
Hand keypoint classification |
29 |
CVI_TDL_SUPPORTED_MODEL _WPODNET |
License plate detection |
30 |
CVI_TDL_SUPPORTED_MODEL _LPRNET_TW |
Taiwan regional license plate recognition |
31 |
CVI_TDL_SUPPORTED_MODEL _LPRNET_CN |
Mercy license plate recognition in mainland China |
32 |
CVI_TDL_SUPPORTED_MODEL _IRLIVENESS |
Live detection |
33 |
CVI_TDL_SUPPORTED_MODEL _YOLOV8POSE |
Gesture |
34 |
CVI_TDL_SUPPORTED_MODEL _SIMCC_POSE |
Gesture |
35 |
CVI_TDL_SUPPORTED_MODEL _YOLOV8_SEG |
80 types of object detection |
36 |
CVI_TDL_SUPPORTED_MODEL _TOPFORMER_SEG |
Semantic segmentation |
37 |
CVI_TDL_SUPPORTED_MODEL _FACELANDMARKER |
Face keypoint detection |
38 |
CVI_TDL_SUPPORTED_MODEL _EYECLASSIFICATION |
Close eye detection |
39 |
CVI_TDL_SUPPORTED_MODEL _YAWNCLASSIFICATION |
Yawning classification |
40 |
CVI_TDL_SUPPORTED_MODEL _SOUNDCLASSIFICATION |
Baby crying recognition |
Here is a table of model IDs, corresponding model files, and the inference functions used:
Index |
Inference Function |
Model file |
---|---|---|
0 |
CVI_TDL_Detection |
yolov3.cvimodel |
1 |
CVI_TDL_Detection |
yolov5m.cvimodel yolov5s.cvimodel |
2 |
CVI_TDL_Detection |
yolov6m.cvimodel yolov6s.cvimodel |
3 |
CVI_TDL_Detection |
yolov7-tiny.cvimodel |
4 |
CVI_TDL_Detection |
yolov8n.cvimodel yolov8s.cvimodel |
5 |
CVI_TDL_Detection |
yolox_m.cvimodel yolox_s.cvimodel |
6 |
CVI_TDL_Detection |
ppyoloe.cvimodel |
7 |
CVI_TDL_Detection |
yolov10n.cvimodel |
8 |
CVI_TDL_Detection |
hardhat_detection_v2.cvimodel |
9 |
CVI_TDL_Detection |
hand_det_qat_640x384.cvimodel |
10 |
CVI_TDL_Detection |
pet_det_640x384.cvimodel |
11 |
CVI_TDL_Detection |
yolov8n_384_640_person _vehicle.cvimodel |
12 |
CVI_TDL_Detection |
meeting_det_640x384.cvimodel |
13 |
CVI_TDL_Detection |
yolov8n_headperson.cvimodel |
14 |
CVI_TDL_Detection |
mobiledetv2-d0-ls.cvimodel mobiledetv2-d1-ls.cvimodel mobiledetv2-d2-ls.cvimodel |
15 |
CVI_TDL_Detection |
mobiledetv2-vehicle-d0-ls.cvimodel |
16 |
CVI_TDL_Detection |
mobiledetv2-pedestrian-d0-ls-384.cvimodel mobiledetv2-pedestrian-d0-ls-448.cvimodel mobiledetv2-pedestrian-d0-ls-640.cvimodel mobiledetv2-pedestrian-d0-ls-768.cvimodel mobiledetv2-pedestrian-d1-ls.cvimodel mobiledetv2-pedestrian-d1-ls-1024.cvimodel |
17 |
CVI_TDL_FaceDetection |
retinaface_mnet0.25_342_608.cvimodel retinaface_mnet0.25_608_342.cvimodel retinaface_mnet0.25_608.cvimodel |
18 |
CVI_TDL_FaceDetection |
scrfd_320_256_ir.cvimodel scrfd_480_270_int8.cvimodel scrfd_480_360_int8.cvimodel scrfd_500m_bnkps_432_768.cvimodel |
19 |
CVI_TDL_FaceDetection |
retinafaceIR_mnet0.25_342_608.cvimodel retinafaceIR_mnet0.25_608_342.cvimodel retinafaceIR_mnet0.25_608_608.cvimodel |
20 |
CVI_TDL_FaceDetection |
retinaface_yolox_fdmask.cvimodel |
21 |
CVI_TDL_FaceLandmarkerDet2 |
pipnet_blurness_v5_64_retinaface _50ep.cvimodel |
22 |
CVI_TDL_FaceQuality |
fqnet-v5_shufflenetv2-softmax.cvimodel |
23 |
CVI_TDL_MaskClassification |
mask_classifier.cvimodel |
24 |
CVI_TDL_FaceRecognition |
cviface-v5-m.cvimodel cviface-v6-s.cvimodel |
25 |
CVI_TDL_FaceAttribute_cls |
face_attr_112_112.cvimodel |
26 |
CVI_TDL_HandClassification |
hand_cls_128x128.cvimodel |
27 |
CVI_TDL_HandKeypoint |
hand_kpt_128x128.cvimodel |
28 |
CVI_TDL_HandKeypointClassification |
hand_kpt_cls9.cvimodel |
29 |
CVI_TDL_LicensePlateDetection |
wpodnet_v0_bf16.cvimodel |
30 |
CVI_TDL_LicensePlateRecognition_TW |
lprnet_v0_tw_bf16.cvimodel |
31 |
CVI_TDL_LicensePlateRecognition_CN |
lprnet_v1_cn_bf16.cvimodel |
32 |
CVI_TDL_IrLiveness |
ir_liveness.cvimodel |
33 |
CVI_TDL_PoseDetection |
yolov8n_pose_384_640.cvimodel |
34 |
CVI_TDL_PoseDetection |
simcc_mv2_pose.cvimodel |
35 |
CVI_TDL_YoloV8_Seg |
yolov8n_seg.cvimodel |
36 |
CVI_TDL_Topformer_Seg |
topformer_tiny.cvimodel |
37 |
CVI_TDL_FaceLandmarker |
face_landmark_bf16.cvimodel |
38 |
CVI_TDL_EyeClassification |
eye_v1_bf16.cvimodel |
39 |
CVI_TDL_YawnClassification |
yawn_v1_bf16.cvimodel |
40 |
CVI_TDL_SoundClassification |
c10_lightv2_mse40_mix.cvimodel |
6.1.2. cvtdl_obj_class_id_e¶
【Description】
This enum defines the categories of object detection, with each category belonging to a group.
Category |
Category Group |
---|---|
CVI_TDL_DET_TYPE_PERSON |
CVI_TDL_DET_GROUP_PERSON |
CVI_TDL_DET_TYPE_BICYCLE |
CVI_TDL_DET_GROUP_VEHICLE |
CVI_TDL_DET_TYPE_CAR |
|
CVI_TDL_DET_TYPE_MOTORBIKE |
|
CVI_TDL_DET_TYPE_AEROPLANE |
|
CVI_TDL_DET_TYPE_BUS |
|
CVI_TDL_DET_TYPE_TRAIN |
|
CVI_TDL_DET_TYPE_TRUCK |
|
CVI_TDL_DET_TYPE_BOAT |
|
CVI_TDL_DET_TYPE_TRAFFIC_LIGHT |
CVI_TDL_DET_GROUP _OUTDOOR |
CVI_TDL_DET_TYPE_FIRE_HYDRANT |
|
CVI_TDL_DET_TYPE_STREET_SIGN |
|
CVI_TDL_DET_TYPE_STOP_SIGN |
|
CVI_TDL_DET_TYPE_PARKING_METER |
|
CVI_TDL_DET_TYPE_BENCH |
|
CVI_TDL_DET_TYPE_BIRD |
CVI_TDL_DET_GROUP_ANIMAL |
CVI_TDL_DET_TYPE_CAT |
|
CVI_TDL_DET_TYPE_DOG |
|
CVI_TDL_DET_TYPE_HORSE |
|
CVI_TDL_DET_TYPE_SHEEP |
|
CVI_TDL_DET_TYPE_COW |
|
CVI_TDL_DET_TYPE_ELEPHANT |
|
CVI_TDL_DET_TYPE_BEAR |
|
CVI_TDL_DET_TYPE_ZEBRA |
|
CVI_TDL_DET_TYPE_GIRAFFE |
|
CVI_TDL_DET_TYPE_HAT |
CVI_TDL_DET_GROUP _ACCESSORY |
CVI_TDL_DET_TYPE_BACKPACK |
|
CVI_TDL_DET_TYPE_UMBRELLA |
|
CVI_TDL_DET_TYPE_SHOE |
|
CVI_TDL_DET_TYPE_EYE_GLASSES |
|
CVI_TDL_DET_TYPE_HANDBAG |
|
CVI_TDL_DET_TYPE_TIE |
|
CVI_TDL_DET_TYPE_SUITCASE |
|
CVI_TDL_DET_TYPE_FRISBEE |
CVI_TDL_DET_GROUP_SPORTS |
CVI_TDL_DET_TYPE_SKIS |
|
CVI_TDL_DET_TYPE_SNOWBOARD |
|
CVI_TDL_DET_TYPE_SPORTS_BALL |
|
CVI_TDL_DET_TYPE_KITE |
|
CVI_TDL_DET_TYPE_BASEBALL_BAT |
|
CVI_TDL_DET_TYPE_BASEBALL_GLOVE |
|
CVI_TDL_DET_TYPE_SKATEBOARD |
|
CVI_TDL_DET_TYPE_SURFBOARD |
|
CVI_TDL_DET_TYPE_TENNIS_RACKET |
|
CVI_TDL_DET_TYPE_BOTTLE |
CVI_TDL_DET_GROUP_KITCHEN |
CVI_TDL_DET_TYPE_PLATE |
|
CVI_TDL_DET_TYPE_WINE_GLASS |
|
CVI_TDL_DET_TYPE_CUP |
|
CVI_TDL_DET_TYPE_FORK |
|
CVI_TDL_DET_TYPE_KNIFE |
|
CVI_TDL_DET_TYPE_SPOON |
|
CVI_TDL_DET_TYPE_BOWL |
|
CVI_TDL_DET_TYPE_BANANA |
CVI_TDL_DET_GROUP_FOOD |
CVI_TDL_DET_TYPE_APPLE |
|
CVI_TDL_DET_TYPE_SANDWICH |
|
CVI_TDL_DET_TYPE_ORANGE |
|
CVI_TDL_DET_TYPE_BROCCOLI |
|
CVI_TDL_DET_TYPE_CARROT |
|
CVI_TDL_DET_TYPE_HOT_DOG |
|
CVI_TDL_DET_TYPE_PIZZA |
|
CVI_TDL_DET_TYPE_DONUT |
|
CVI_TDL_DET_TYPE_CAKE |
|
CVI_TDL_DET_TYPE_CHAIR |
CVI_TDL_DET_GROUP _FURNITURE |
CVI_TDL_DET_TYPE_SOFA |
|
CVI_TDL_DET_TYPE_POTTED_PLANT |
|
CVI_TDL_DET_TYPE_BED |
|
CVI_TDL_DET_TYPE_MIRROR |
|
CVI_TDL_DET_TYPE_DINING_TABLE |
|
CVI_TDL_DET_TYPE_WINDOW |
|
CVI_TDL_DET_TYPE_DESK |
|
CVI_TDL_DET_TYPE_TOILET |
|
CVI_TDL_DET_TYPE_DOOR |
|
CVI_TDL_DET_TYPE_TV_MONITOR |
CVI_TDL_DET_GROUP _ELECTRONIC |
CVI_TDL_DET_TYPE_LAPTOP |
|
CVI_TDL_DET_TYPE_MOUSE |
|
CVI_TDL_DET_TYPE_REMOTE |
|
CVI_TDL_DET_TYPE_KEYBOARD |
|
CVI_TDL_DET_TYPE_CELL_PHONE |
|
CVI_TDL_DET_TYPE_MICROWAVE |
CVI_TDL_DET_GROUP _APPLIANCE |
CVI_TDL_DET_TYPE_OVEN |
|
CVI_TDL_DET_TYPE_TOASTER |
|
CVI_TDL_DET_TYPE_SINK |
|
CVI_TDL_DET_TYPE_REFRIGERATOR |
|
CVI_TDL_DET_TYPE_BLENDER |
|
CVI_TDL_DET_TYPE_BOOK |
CVI_TDL_DET_GROUP_INDOOR |
CVI_TDL_DET_TYPE_CLOCK |
|
CVI_TDL_DET_TYPE_VASE |
|
CVI_TDL_DET_TYPE_SCISSORS |
|
CVI_TDL_DET_TYPE_TEDDY_BEAR |
|
CVI_TDL_DET_TYPE_HAIR_DRIER |
|
CVI_TDL_DET_TYPE_TOOTHBRUSH |
|
CVI_TDL_DET_TYPE_HAIR_BRUSH |
6.1.3. cvtdl_obj_det_group_type_e¶
【Description】
This enum defines the object category groups.
Category Group |
Description |
---|---|
CVI_TDL_DET_GROUP_ALL |
Full class |
CVI_TDL_DET_GROUP_PERSON |
Human form |
CVI_TDL_DET_GROUP_VEHICLE |
Means of transportation |
CVI_TDL_DET_GROUP_OUTDOOR |
Outdoors |
CVI_TDL_DET_GROUP_ANIMAL |
Animal |
CVI_TDL_DET_GROUP_ACCESSORY |
Accessories |
CVI_TDL_DET_GROUP_SPORTS |
Movement |
CVI_TDL_DET_GROUP_KITCHEN |
Kitchen |
CVI_TDL_DET_GROUP_FOOD |
Food |
CVI_TDL_DET_GROUP_FURNITURE |
Furniture |
CVI_TDL_DET_GROUP_ELECTRONIC |
Electronic equipment |
CVI_TDL_DET_GROUP_APPLIANCE |
Appliance |
CVI_TDL_DET_GROUP_INDOOR |
Indoor articles |
CVI_TDL_DET_GROUP_MASK_HEAD |
Custom category |
CVI_TDL_DET_GROUP_MASK_START |
Custom category start |
CVI_TDL_DET_GROUP_MASK_END |
Custom category end |
6.1.4. feature_type_e¶
【enum】
Value |
Parameter |
Description |
---|---|---|
0 |
TYPE_INT8 |
int8_t feature type |
1 |
TYPE_UINT8 |
uint8_t feature type |
2 |
TYPE_INT16 |
int16_t feature type |
3 |
TYPE_UINT16 |
uint16_t feature type |
4 |
TYPE_INT32 |
int32_t feature type |
5 |
TYPE_UINT32 |
uint32_t feature type |
6 |
TYPE_BF16 |
bf16 feature type |
7 |
TYPE_FLOAT |
float feature type |
6.1.5. meta_rescale_type_e¶
【enum】
Value |
Parameter |
Description |
---|---|---|
0 |
RESCALE_UNKNOWN |
Unknown |
1 |
RESCALE_NOASPECT |
Direct adjustment without proportional scaling |
2 |
RESCALE_CENTER |
Padding around all sides |
3 |
RESCALE_RB |
Padding on the bottom right side |
6.1.6. cvtdl_bbox_t¶
Data Type |
Parameter |
Description |
---|---|---|
float |
x1 |
The x value of the coordinate at the top left of the detection frame |
float |
y1 |
The y value of the coordinate on the upper left of the detection frame |
float |
x2 |
The x value of the lower right coordinate of the detection box |
float |
y2 |
The y value of the lower right coordinate of the detection box |
float |
score |
The confidence level of the detection box |
6.1.7. cvtdl_feature_t¶
Data Type |
Parameter |
Description |
---|---|---|
int8_t* |
ptr |
Address |
uint32_t |
size |
Characteristic dimension |
feature_type_e |
type |
Characteristic pattern |
6.1.8. cvtdl_pts_t¶
Data Type |
Parameter |
Description |
---|---|---|
float* |
x |
Coordinate x |
float* |
y |
Coordinate y |
uint32_t |
size |
Number of coordinate points |
6.1.9. cvtdl_4_pts_t¶
Data Type |
Parameter |
Description |
---|---|---|
float |
x[4] |
The x-coordinate values of the four coordinate points |
float |
y[4] |
The y coordinate values of the four coordinate points |
6.1.10. cvtdl_vpssconfig_t¶
Data Type |
Parameter |
Description |
---|---|---|
VPSS_SCALE_COEF_E |
chn_coeff |
Rescale mode |
VPSS_CHN_ATTR_S |
chn_attr |
VPSS attribute data |
6.1.11. cvtdl_tracker_t¶
Data Type |
Parameter |
Description |
---|---|---|
uint32_t |
size |
Number of trace messages |
cvtdl_tracker_info_t* |
info |
Trace message structure |
6.1.12. cvtdl_tracker_info_t¶
Data Type |
Parameter |
Description |
---|---|---|
cvtdl_trk_state_type_t |
state |
Tracking state |
cvtdl_bbox_t |
bbox |
Tracking the Bounding Box of predictions |
6.1.13. cvtdl_trk_state_type_t¶
【enum】
Value |
Parameter |
Description |
---|---|---|
0 |
CVI_TRACKER_NEW |
The tracking status is new |
1 |
CVI_TRACKER_UNSTABLE |
The tracking state is unstable |
2 |
CVI_TRACKER_STABLE |
The tracking state is stable |
6.1.14. cvtdl_deepsort_config_t¶
Data Type |
Parameter |
Description |
---|---|---|
float |
max_distance_iou |
Maximum IOU distance for BBox matching |
float |
ma x_distance_consine |
Maximum consine distance during Feature matching |
int |
max_unmatched_times _for_bbox_matching |
The number of maximum unmatched times of the target participating in BBox matching |
bool |
enable_internal_FQ |
Enable internal feature quality |
cvtdl_kalman_filter_config_t |
kfilter_conf |
Kalman filter settings |
cvtdl_kalman_tracker_config_t |
ktracker_conf |
Kalman tracker settings |
6.1.15. cvtdl_kalman_filter_config_t¶
Data Type |
Parameter |
Description |
---|---|---|
bool |
enable_X_constraint_0 |
Enable 0th X constraint |
bool |
enable_X_constraint_1 |
Enable 1st X constraint |
float |
X_constraint_min[8] |
X Constraint lower bound |
float |
X_constraint_max[8] |
X Constraint upper limit |
bool |
enable_bounding_stay |
Keep boundaries |
mahalanobis _confidence_e |
confidence_level |
Mahalanobis distance confidence |
float |
chi2_threshold |
Chi-square threshold |
float |
Q_std_alpha[8] |
Process Noise parameter |
float |
Q_std_beta[8] |
Process Noise parameter |
int |
Q_std_x_idx[8] |
Process Noise parameter |
float |
R_std_alpha[4] |
Measurement Noise parameter |
float |
R_std_beta[4] |
Measurement Noise parameter |
int |
R_std_x_idx[4] |
Measurement Noise parameter |
【Description】
For the motion state X of a tracked target…
Process Nose, Q, where
\(Q\lbrack i\rbrack = \left( {Alpha}_{Q}\lbrack i\rbrack \bullet X\left\lbrack {Idx}_{Q}\lbrack i\rbrack \right\rbrack + {Beta}_{Q}\lbrack i\rbrack \right)^{2}\)
Measurement Nose, R, The formula for motion deviation is the same as follows:
6.1.16. cvtdl_kalman_tracker_config_t¶
Data Type |
Parameter |
Description |
---|---|---|
int |
max_unmatched_num |
Track the maximum number of lost targets |
int |
accreditation_threshold |
The threshold at which the tracking state changes to stable |
int |
feature_budget_size |
Save the maximum number of tracking target features |
int |
feature_update_interval |
Time interval of updating feature |
bool |
enable_QA_feature_init |
Enable QA feature initialization |
bool |
enable_QA_feature_update |
Enable QA feature updates |
float |
feature_init_quality_threshold |
Feature initialization quality threshold |
float |
feature_update_quality_threshold |
Feature update quality threshold |
float |
P_std_alpha[8] |
Initial Covariance parameter |
float |
P_std_beta[8] |
Initial Covariance parameter |
int |
P_std_x_idx[8] |
Initial Covariance parameter |
【Description】
Initial Covariance, P, The formula for updating the initial covariance is similar to that of the process noise covariance:
6.1.17. cvtdl_liveness_ir_position_e¶
【enum】
Value |
Parameter |
Description |
---|---|---|
0 |
LIVENESS_IR_LEFT |
The IR lens is to the left of the RGB lens |
1 |
LIVENESS_IR_RIGHT |
The IR lens is to the right of the RGB lens |
6.1.18. cvtdl_head_pose_t¶
Data Type |
Parameter |
Description |
---|---|---|
float |
yaw |
Yaw angle |
float |
pitch |
Pitch angle |
float |
roll |
Angle of roll |
float |
facialUnitNormalVector[3] |
The face is oriented towards the bearing |
6.1.19. cvtdl_face_info_t¶
Data Type |
Parameter |
Description |
---|---|---|
char |
name[128] |
Face name |
uint64_t |
unique_id |
Face ID |
cvtdl_bbox_t |
bbox |
Face detection frame |
cvtdl_pts_t |
pts |
Face feature point |
cvtdl_feature_t |
feature |
Facial features |
cvtdl_face_emotion_e |
emotion |
Expression |
cvtdl_face_gender_e |
gender |
Gender |
cvtdl_face_race_e |
race |
Race |
float |
score |
Score |
float |
age |
Age |
float |
liveness_score |
Live probability |
float |
hardhat_score |
Hard hat score |
float |
mask_score |
Probability of wearing face mask |
float |
recog_score |
Recognition score |
float |
face_quality |
Human face quality |
float |
pose_score |
Pose score |
float |
pose_score1 |
Pose score 1 |
float |
sharpness_score |
Sharpness score |
float |
blurness |
Blurness |
cvtdl_head_pose_t |
head_pose |
Facial angle information |
int |
track_state |
Track state |
6.1.20. cvtdl_face_t¶
Data Type |
Parameter |
Description |
---|---|---|
uint32_t |
size |
Number of faces |
uint32_t |
width |
The width of the original picture |
uint32_t |
height |
The height of the original picture |
meta_rescale_type_e* |
rescale_type |
Rescale type |
cvtdl_face_info_t* |
info |
Face synthesis information |
cvtdl_dms_t* |
dms |
Driver monitoring system |
6.1.21. cvtdl_pose17_meta_t¶
Data Type |
Parameter |
Description |
---|---|---|
float |
x[17] |
x-coordinate of 17 skeletal keypoints |
float |
y[17] |
y-coordinate of 17 skeletal keypoints |
float |
score[17] |
Predicted confidence value of 17 skeletal keypoints |
6.1.22. cvtdl_vehicle_meta¶
Data Type |
Parameter |
Description |
---|---|---|
cvtdl_4_pts_t |
license_pts |
Coordinates of the 4 corners of the license plate |
cvtdl_bbox_t |
license_bbox |
Bounding box of the license plate |
char[125] |
license_char |
License plate number |
【Description】
The four corner coordinates of the license plate are upper left, upper right, lower right to lower left in order.
6.1.23. cvtdl_class_filter_t¶
Data Type |
Parameter |
Description |
---|---|---|
uint32_t* |
preserved_class_ids |
The category id to keep |
uint32_t |
num_preserved_classes |
The number of category ids to keep |
6.1.24. cvtdl_dms_t¶
Data Type |
Parameter |
Description |
---|---|---|
float |
reye_score |
Right eye opening and closing score |
float |
leye_score |
Left eye opening and closing score |
float |
yawn_score |
Mouth closure score |
float |
phone_score |
Talking on the phone score |
float |
smoke_score |
Smoking score |
cvtdl_pts_t |
landmarks_106 |
106 feature points |
cvtdl_pts_t |
landmarks_5 |
Five feature points |
cvtdl_head_pose_t |
head_pose |
Face angle calculated from 106 feature points |
cvtdl_dms_od_t |
dms_od |
Object detection results in the vehicle |
6.1.25. cvtdl_dms_od_t¶
Data Type |
Parameter |
Description |
---|---|---|
uint32_t |
size |
There are a couple of things |
uint32_t |
width |
Width |
uint32_t |
height |
Height |
meta_rescale_type_e |
rescale_type |
The shape of rescale |
cvtdl_dms_od_info_t* |
info |
Information about objects |
6.1.26. cvtdl_dms_od_info_t¶
Data Type |
Parameter |
Description |
---|---|---|
char[128] |
name |
Name of object |
int |
classes |
Category of objects |
cvtdl_bbox_t |
bbox |
Object bounding Box |
6.1.27. cvtdl_face_emotion_e¶
【Description】
Emotion
Emotion |
Description |
---|---|
EMOTION_UNKNOWN |
unknown |
EMOTION_HAPPY |
Be happy |
EMOTION_SURPRISE |
Surprise |
EMOTION_FEAR |
Fear of |
EMOTION_DISGUST |
Aversion to |
EMOTION_SAD |
Be sad |
EMOTION_ANGER |
Get angry |
EMOTION_NEUTRAL |
Nature |
6.1.28. cvtdl_face_race_e¶
Race |
Description |
---|---|
RACE_UNKNOWN |
Unknown |
RACE_CAUCASIAN |
The Caucasian |
RACE_BLACK |
Black people |
RACE_ASIAN |
The Asians |
6.1.29. cvtdl_adas_meta¶
Data Type |
Parameter |
Description |
---|---|---|
float |
dis |
distance |
float |
speed |
speed |
adas_state_e |
state |
adas attribute, enum NORMAL, START, COLLISION_WARNING, DANGER four stages |
6.1.30. cvtdl_mask_meta¶
Data Type |
Parameter |
Description |
---|---|---|
uint8_t |
mask |
Mask feature |
float |
mask_point |
Key point of mask contour |
uint32_t |
mask_point_size |
Number of critical points of mask contour |
6.1.31. cvtdl_pedestrian_meta¶
Data Type |
Parameter |
Description |
---|---|---|
cvtdl_pose17_meta_t |
pose17 |
Human 17 key points |
bool |
fall |
Whether to fall |
6.1.32. cvtdl_object_info_t¶
Data Type |
Parameter |
Description |
---|---|---|
char |
name |
Object class name |
uint64_t |
unique_id |
Unique id |
cvtdl_box_t |
bbox |
Bounding box |
cvtdl_feature_t |
feature |
Feature of objects |
int |
classes |
ID of category |
cvtdl_mask_meta |
mask_properity |
Feature mask attribute |
cvtdl_vehicle_meta |
vehicle_property |
Property of vehicle |
cvtdl_pedestrian_meta |
pedestrian_property |
Attributes of pedestrians |
cvtdl_adas_meta |
adas_properity |
ADAS attribute |
int |
track_state |
Track state |
6.1.33. cvtdl_object_t¶
Data Type |
Parameter |
Description |
---|---|---|
uint32_t |
size |
The number of objects in info |
uint32_t |
width |
The width of the original image |
uint32_t |
height |
The height of the original picture |
uint32_t |
mask_height |
The height of the feature mask |
uint32_t |
mask_width |
The weight of the feature mask |
uint32_t |
entry_num |
Number of entries |
uint32_t |
miss_num |
Number of miss |
meta_rescale_type_e |
rescale_type |
resize method used in model pre-processing |
cvtdl_object_info_t* |
info |
Object information |
6.1.34. cvtdl_handpose21_meta_t¶
Data Type |
Parameter |
Description |
---|---|---|
float |
xn[21] |
normalize x points |
float |
x[21] |
x points |
float |
yn[21] |
normalize y points |
float |
y[21] |
y points |
float |
bbox_x |
the x-coordinate of the box |
float |
bbox_y |
the y-coordinate of the box |
float |
bbox_w |
The width of box |
float |
bbox_h |
The height of box |
int |
label |
Gesture category |
float |
score |
Gesture score |
6.1.35. cvtdl_handpose21_meta_ts¶
Data Type |
Parameter |
Description |
---|---|---|
uint32_t |
size |
The number of hands detected |
uint32_t |
width |
The width of image |
uint32_t |
height |
The height of image |
cvtdl_handpose21_meta_t* |
info |
Hand keypoints |
6.1.36. cvtdl_seg_t¶
Data Type |
Parameter |
Description |
---|---|---|
char |
img_name[128] |
input image name |
int |
srcWidth |
input image width |
int |
srcHeight |
input image high |
uint8_t |
class_id |
Feature category ID |
uint8_t |
class_conf |
feature category confidence |
6.1.37. Yolov5PreParam¶
Data Type |
Parameter |
Description |
---|---|---|
float |
factor[3] |
Scaling factor |
float |
mean[3] |
Image mean |
meta_rescale_type_e |
rescale_type |
Scaling mode |
bool* |
pad_reverse |
Reverse padding |
bool* |
keep_aspect_ratio |
Keep aspect ratio |
bool* |
use_quantize_scale |
Qantization scaling |
bool* |
use_crop |
Crop to resize an image |
VPSS_SCALE_COEF_E* |
resize_method |
Scaling method |
PIXEL_FORMAT_E* |
format |
Image format |
6.1.38. YOLOV5AlgParam¶
Data Type |
Parameter |
Description |
---|---|---|
uint32_t |
anchors[3][3][2] |
model anchors |
float |
conf_thresh |
confidence threshold |
float |
nms_thresh |
RMS Threshold |
6.2. CVI_TDL_Service¶
6.2.1. cvtdl_service_feature_matching_e¶
【Description】
The method used for feature comparison calculation, currently only supports Cosine Similarity.
【Member】
Parameter |
Description |
---|---|
COS_SIMILARITY |
Cosine similarity |
6.2.2. cvtdl_service_feature_array_t¶
【Description】
Feature array, this structure contains the pointer to the feature array, length, number of features, and feature type information.
This structure needs to be passed in when registering the feature library.
【Member】
Data Type |
Parameter |
Description |
---|---|---|
int8_t* |
ptr |
Feature array pointer |
uint32_t |
feature_length |
Single characteristic length |
uint32_t |
data_num |
Number of features |
feature_type_e |
type |
Type of feature |
6.2.3. cvtdl_service_brush_t¶
【Description】
Brush structure for drawing, which specifies the RGB color and brush size to be used.
【Member】
Data Type |
Parameter |
Description |
---|---|---|
Inner structure |
color |
The RGB value to use |
uint32_t |
size |
Brush size |
6.2.4. cvtdl_area_detect_e¶
【enum】
Data Type |
Parameter |
Description |
---|---|---|
0 |
UNKNOWN |
The int8_t feature type |
1 |
NO_INTERSECT |
Not intersect |
2 |
ON_LINE |
On line |
3 |
CROSS_LINE_POS |
Forward cross |
4 |
CROSS_LINE_NEG |
Negative cross |
5 |
INSIDE_POLYGON |
Inside the polygon |
6 |
OUTSIDE_POLYGON |
Outside the polygon |