darcyai.perceptor.coral.people_perceptor

PeoplePerceptor Objects

class PeoplePerceptor(CoralPerceptorBase, PeoplePerceptorBase)

Perceptor for detecting people in an image.

Perceptor Config:

minimum_face_threshold (float):
    Confidence threshold for detecting the face as a percent certainty
    Default value: 0.4

minimum_body_threshold (float):
    Confidence threshold for detecting the body as a percent certainty
    Default value: 0.2

minimum_face_height (int):
    The minimum height of the persons face in pixels that we want to work with
    Default value: 20

minimum_body_height (int):
    The minimum height of the persons body in pixels that we want to work with
    Default value: 120

show_pose_landmark_dots (bool):
    Show pose landmark dots (nose, ears, elbows, etc.)
    Default value: False

show_body_rectangle (bool):
    Draw a rectangle around the persons body
    Default value: False

show_face_rectangle (bool):
    Draw a rectangle around the persons face
    Default value: False

face_rectangle_color (rgb):
    Color of the face rectangle
    Default value: RGB(255, 0, 0)

face_rectangle_thickness (int):
    Thickness of the face rectangle
    Default value: 1

body_rectangle_color (rgb):
    Color of the body rectangle
    Default value: RGB(0, 255, 0)

body_rectangle_thickness (int):
    Thickness of the body rectangle
    Default value: 1

pose_landmark_dot_confidence_threshold (float):
    Confidence threshold for identifying pose landmarks as a percent certainty
    Default value: 0.5

pose_landmark_dot_size (int):
    Size of the pose landmark dots
    Default value: 1

pose_landmark_dot_color (rgb):
    Color of the pose landmark dots
    Default value: RGB(255, 255, 255)

show_face_position_arrow (bool):
    Show the arrow that indicates which direction the person is looking
    Default value: False

face_position_arrow_color (rgb):
    Color of the face position arrow
    Default value: RGB(255, 255, 255)

face_position_arrow_stroke (int):
    Thickness of the face position arrow
    Default value: 1

face_position_arrow_offset_x (int):
    X offset for the face position arrow
    Default value: 0

face_position_arrow_offset_y (int):
    Y offset for the face position arrow
    Default value: -30

face_position_arrow_length (int):
    Length of the face position arrow
    Default value: 20

face_position_left_right_threshold (float):
    Threshold for detecting if the person is looking left or right
    Default value: 0.3

face_position_straight_threshold (float):
    Threshold for detecting if the person is looking straight
    Default value: 0.7

show_forehead_center_dot (bool):
    Show a dot in the center of the persons forehead
    Default value: False

forehead_center_dot_color (rgb):
    Color of the forehead center dot
    Default value: RGB(255, 255, 255)

forehead_center_dot_size (int):
    Size of the forehead center dot
    Default value: 1

face_rectangle_y_factor (float):
    Size adjustment factor for the height of the persons face, which can be used to make sure objects like hair and hats are captured
    Default value: 1.0

show_centroid_dots (bool):
    Show centroid information (center of the face or body)
    Default value: False

centroid_dots_color (rgb):
    Color of the centroid information
    Default value: RGB(255, 255, 255)

centroid_dots_size (int):
    Size of the centroid information
    Default value: 1

object_tracking_allowed_missed_frames (int):
    Object tracking allowed missed frames
    Default value: 50

object_tracking_color_sample_pixels (int):
    Number of pixels to use for color sampling when tracking objects
    Default value: 4

object_tracking_info_history_count (int):
    Number of video frames used to track an object in the field of view
    Default value: 3

object_tracking_removal_count (int):
    Number of frames to wait before removing an object from tracking
    Default value: 50

object_tracking_centroid_weight (float):
    Level of importance that centroid data has when tracking objects
    Default value: 0.25

object_tracking_color_weight (float):
    Level of importance that color data has when tracking objects
    Default value: 0.25

object_tracking_vector_weight (float):
    Level of importance that vector data has when tracking objects
    Default value: 0.25

object_tracking_size_weight (float):
    Level of importance that size data has when tracking objects
    Default value: 0.25

object_tracking_creation_m (int):
    Minimum number of frames out of N frames that an object must be present in the field of view before it is tracked
    Default value: 10

object_tracking_creation_n (int):
    Total number of frames used to evaluate an object before it is tracked
    Default value: 7

person_tracking_creation_m (int):
    Minimum number of frames out of N frames needed to promote a tracked object to a person
    Default value: 20

person_tracking_creation_m (int):
    Total number of frames used to evaluate a tracked object before it is promoted to a person
    Default value: 16

show_person_id (bool):
    Show anonymous unique identifier for the person
    Default value: False

person_data_line_color (rgb):
    Color of the person data line
    Default value: RGB(255, 255, 255)

person_data_line_thickness (int):
    Thickness of the person data line
    Default value: 1

person_data_identity_text_color (rgb):
    Color of the person data identity text
    Default value: RGB(255, 255, 255)

person_data_identity_text_stroke (int):
    Thickness of the person data identity text
    Default value: 1

person_data_identity_text_font_size (float):
    Font size of the person data identity text
    Default value: 1.0

person_data_text_offset_x (int):
    X offset of the person data text
    Default value: 30

person_data_text_offset_y (int):
    Y offset of the person data text
    Default value: -40

identity_text_prefix (str):
    Text information that you want to display at the beginning of the person ID
    Default value: Person ID:

rolling_video_storage_frame_count (int):
    Number of the video frames to store while processing
    Default value: 100

Events:

new_person_entered_scene
person_facing_new_direction
new_person_in_front
person_left_scene
identity_determined_for_person
person_got_too_close
person_went_too_far_away
max_person_count_reached
person_count_fell_below_maximum
person_occluded