Result | @vladmandic/human

Result interface definition for Human library

Contains all possible detection results

Hierarchy

Result

Index

Properties

body

body: { id: number; part: string; position: { x: number; y: number; z: number }; presence: number; score: number }[]

Body results

Array of individual results with one object per detected body Each results has:

body id number
body part name
part position with x,y,z coordinates
body part score value
body part presence value

canvas

canvas: HTMLCanvasElement | OffscreenCanvas

face

face: { age: number; annotations: { part: string; points: [number, number, number][][] }[]; box: [number, number, number, number]; boxConfidence: number; boxRaw: [number, number, number, number]; confidence: number; embedding: number[]; emotion: { emotion: string; score: number }[]; faceConfidence: number; gender: string; genderConfidence: number; iris: number; mesh: [number, number, number][]; meshRaw: [number, number, number][]; rotation: { angle: { pitch: number; roll: number; yaw: number }; matrix: [number, number, number, number, number, number, number, number, number][] }; tensor: any }[]

Face results Combined results of face detector, face mesh, age, gender, emotion, embedding, iris models Some values may be null if specific model is not enabled

Array of individual results with one object per detected face Each result has:

overal detection confidence value
box detection confidence value
mesh detection confidence value
box as array of [x, y, width, height], normalized to image resolution
boxRaw as array of [x, y, width, height], normalized to range 0..1
mesh as array of [x, y, z] points of face mesh, normalized to image resolution
meshRaw as array of [x, y, z] points of face mesh, normalized to range 0..1
annotations as array of annotated face mesh points
age as value
gender as value
genderConfidence as value
emotion as array of possible emotions with their individual scores
iris as distance value
angle as object with values for roll, yaw and pitch angles
tensor as Tensor object which contains detected face

gesture

gesture: ({ face: number; gesture: string } | { gesture: string; iris: number } | { body: number; gesture: string } | { gesture: string; hand: number })[]

Gesture results

Array of individual results with one object per detected gesture Each result has:

part: part name and number where gesture was detected: face, iris, body, hand
gesture: gesture detected

hand

hand: { annotations: { part: string; points: [number, number, number][][] }[]; box: [number, number, number, number]; boxRaw: [number, number, number, number]; confidence: number; landmarks: [number, number, number][] }[]

Hand results

Array of individual results with one object per detected hand Each result has:

confidence as value
box as array of [x, y, width, height], normalized to image resolution
boxRaw as array of [x, y, width, height], normalized to range 0..1
landmarks as array of [x, y, z] points of hand, normalized to image resolution
annotations as array of annotated face landmark points

object

object: { box: number[]; boxRaw: number[]; center: number[]; centerRaw: number[]; class: number; label: string; score: number; strideSize: number }[]

Object results

Array of individual results with one object per detected gesture Each result has:

score as value
label as detected class name
center as array of [x, y], normalized to image resolution
centerRaw as array of [x, y], normalized to range 0..1
box as array of [x, y, width, height], normalized to image resolution
boxRaw as array of [x, y, width, height], normalized to range 0..1

performance

performance: { any: any }

Type declaration

any: any