Связаться с намиПодключиться

Vision Classifier API, gRPC: ImageClassifierService.Annotate

Статья создана
Обновлена 17 октября 2024 г.

gRPC request

rpc Annotate (AnnotationRequest) returns (AnnotationResponse)

AnnotationRequest

{
  "image": {
    // Includes only one of the fields `content`
    "content": "bytes",
    // end of the list of possible fields
    "imageType": "ImageType"
  }
}

request for annotation

Field

Description

image

Image

image to annotate

Image

Field

Description

content

bytes

bytes with data

Includes only one of the fields content.

imageType

enum ImageType

type of data

  • IMAGE_TYPE_UNSPECIFIED
  • JPEG
  • PNG

AnnotationResponse

{
  "requestId": "string",
  "classifierSpecification": {
    "labels": [
      {
        "name": "string",
        "description": "string"
      }
    ],
    "classificationType": "ClassificationType"
  },
  "annotations": [
    {
      "label": {
        "name": "string",
        "description": "string"
      },
      "confidence": "double"
    }
  ]
}

Field

Description

requestId

string

internal service requestId

classifierSpecification

ClassifierSpecification

class specification

annotations[]

ClassAnnotation

annotations for each class

ClassifierSpecification

Specification of model used for annotation

Field

Description

labels[]

Label

List of labels, annotated by service

classificationType

enum ClassificationType

type of annotation: exclusive (multi-class) or non-exclusive (multi-label)

  • CLASSIFICATION_TYPE_UNSPECIFIED
  • MULTI_LABEL
  • MULTI_CLASS

Label

Description of single label

Field

Description

name

string

Label name

description

string

human readable description of label

ClassAnnotation

Image annotation for specific label

Field

Description

label

Label

list of annotated labels

confidence

double

confidence for each label