Skip to main content

Machine Learning Data Types

This page describes all the Object types defined and used by Machine Learning Activities.

DateParserInfo

Properties

PropertyTypeDescription
DateDateTimeThe parsed and modified Date value
ConfidenceDoubleThe confidence of the model in the range 0 to 1

NumberParserInfo

Properties

PropertyTypeDescription
NumberlongThe parsed and modified Numeric value
ConfidenceDoubleThe confidence of the model in the range 0 to 1

DocumentInfo

Properties

PropertyTypeDescription
DocumentTypesDictionary<String, Float>List of Document Types and the corresponding confidence
DomainStringThe domain to which this document belongs

SentimentInfo

Properties

PropertyTypeDescription
SentimentStringContains one of the following values: Positive, Negative, or Neutral.
TextStringThe input text.
ConfidenceDoubleThe confidence of the model in the range 0 to 1

LanguageDetectionResult

It represents language detection results obtained from cognitive services.

Properties

PropertyTypeDescription
LanguageStringLanguage name.
LanguageCodeStringShortcode of detected Language in ISO 639-1 format. Refer to List of ISO 639-1 codes
ConfidenceScoreInt32Score between 0 and 100, indicating the confidence for detecting this language.

Entity

It represents an entity identified by cognitive services.

Properties

PropertyTypeDescription
TextStringText of the entity.
ConfidenceScoreInt32Score between 0 to 100, indicating the confidence for detecting this entity.
LengthInt32Length of the entity text.
OffsetInt32Starting position of the entity text within the document text.

NamedEntity

It represents a named entity identified by cognitive services.

Properties

PropertyTypeDescription
TextStringText of the entity.
CategoryStringCategory of the identified entity.
SubCategoryStringSubcategory of the entity.
ConfidenceScoreInt32Score between 0 to 100, indicating the confidence for detecting this entity.
LengthInt32Length of the entity text.
OffsetInt32Starting position of the entity text within the document text.

BoundingBox

Represents an abstract rectangle that surrounds elements on a document or form. It is used as a reference point for object detection.
Each point coordinate represents a corner of the box. For Image files, coordinates are in pixels. For PDF files, coordinates are in inches

Properties

PropertyTypeDescription
TopLeftPointFContains top left coordinates of the bounding box.
TopRightPointFContains top right coordinates of the bounding box.
BottomLeftPointFContains bottom left coordinates of the bounding box.
BottomRightPointFContains bottom right coordinates of the bounding box.

CheckMark

Represents a check mark (selection mark) detected in a form

Properties

PropertyTypeDescription
TextStringText associated with the check box (if any)
ConfidenceScoreFloatConfidence score of this detection in the range of 0 to 100
PageNumberIntegerPage number of detected check mark
CheckStateBooleanIndicates whether the SelectionMark was selected or unselected
BoxBoundingBoxBox coordinates of the check mark detection

FormText

Represents a word inside a line or cell element inside a table cell

Properties

PropertyTypeDescription
PageNumberIntegerPage number of element
ConfidenceScoreFloatConfidence score of the element detection in range 0 to 100
TextStringText detected in the element
BoxBoundingBoxBox coordinates of the element

FormTableCell

Represents a detected table cell in the form

Properties

PropertyTypeDescription
TextStringText in the table cell
ConfidenceScoreFloatConfidence score of the table cell detection in range 0 to 100
BoxBoundingBoxBox coordinates of the table cell
PageNumberIntegerPage number of detected table cell
ColumnIndexIntegerIndex of the column containing the table cell
RowIndexIntegerIndex of the row containing the table cell
ColumnSpanIntegerColumn span of the table cell
RowSpanIntegerRow span of the table cell
HeaderCellBooleanIndicates whether the table cell is in header row
FooterCellBooleanIndicates whether the table cell is in footer row
CellElementsFormTextListing of elements inside the table cell

FormTable

Represents a detected table in a form

Properties

PropertyTypeDescription
PageNumberIntegerPage number of detected table
BoxBoundingBoxBox coordinates of the table
ConfidenceScoreFloatConfidence score of the table detection in range 0 to 100
RowCountIntegerNumber of rows in the table
ColumnCountIntegerNumber of columns in the table
CellsFormTableCellListing of all cells inside the table

FormLine

Represents a detected line of text inside a form

Properties

PropertyTypeDescription
HandwrittenBooleanIndicates whether the line was handwritten or not
WordsFormTextListing of words inside the line
PageNumberIntegerPage number of the detected line
ConfidenceScoreFloatConfidence score of the line detection in the range between 0 to 100
TextStringDetected text in the line
BoxBoundingBoxBox coordinates of the line

FormPage

Represents a page of a form

Properties

PropertyTypeDescription
HeightFloatHeight of the detected page
WidthFloatWidth of the detected page
PageNumberIntegerPage number of the detected form page
TextAngleFloatOrientation of detected text in degrees. Value in the range between -180 to 180
LinesFormLineListing of lines detected in the page
TablesFormTableListing of tables detected in the page
CheckMarksCheckMarkListing of check marks (selection marks) detected on the page

FormField

Represents a field inside the form

Properties

PropertyTypeDescription
ConfidenceScoreFloatConfidence score in range 0 to 100 of the detected form field
LabelStringLabel text of the detected field
LabelBoxBoundingBoxBox coordinates of the field label
LabelPageNumberIntegerPage number of the field label
LabelElementsFormTextElements of the field label
ValueObjectValue of the identified field. Cast to its appropriate type
ValueBoxBoundingBoxBox coordinates of the field value
ValuePageNumberIntegerPage number of the field value
ValueElementsFormTextElements of the field value
ValueTextStringTextual value of the identified field
HasValueBooleanIndicates whether a value is available for this form field

FormFieldValue

Represents a field inside the form

Properties

PropertyTypeDescription
ConfidenceScoreFloatConfidence score in range 0 to 100 of the detected form field
LabelStringLabel text of the detected field
LabelBoxBoundingBoxBox coordinates of the field label
LabelPageNumberIntegerPage number of the field label
LabelElementsFormTextElements of the field label
ValueTValue of the identified field. Cast to its appropriate type
ValueBoxBoundingBoxBox coordinates of the field value
ValuePageNumberIntegerPage number of the field value
ValueElementsFormTextElements of the field value
ValueTextStringTextual value of the identified field

Form

Represents the type of form

Properties

PropertyTypeDescription
FormTypeFormTypeType of the detected form. Can be a BusinessCard, IdentityDocument, Invoice, or Receipt
FormTypeConfidenceScoreFloatConfidence score in the range from 0 to 100 of the detected form type
PagesFormPageListing of pages identified in the form layout
FieldsFormFieldDictionary of fields identified in the form. The field name is used as a key in this dictionary

InvoiceLineItem

Represents a line item inside an invoice

Properties

PropertyTypeDescription
TextStringTextual representation of the line item
AmountFormFieldValue FloatAmount detected in the line item
DescriptionFormFieldValue StringDescription of item detected in the line item
QuantityFormFieldValue FloatQuantity detected in the line item
UnitPriceFormFieldValue FloatPrice per unit detected in the line item
ProductCodeFormFieldValue StringProduct code detected in the line item
UnitFormFieldValue StringUnit of the item. e.g., hours, nos, etc.
DateFormFieldValue DateTimeDate detected in the line item
TaxFormFieldValue FloatTax amount detected in the line item. It could be represented as the actual amount, percentage, etc.
ValueAddedTaxFormFieldValue FloatVAT amount detected in the line item. The field name in Azure: VAT
FieldsFormFieldDictionary of fields identified in the line item. The field name is used as a key in this dictionary

Invoice

Represents an invoice form

Properties

PropertyTypeDescription
CustomerNameFormFieldValue StringInvoiced customer name
CustomerIdFormFieldValue StringCustomer reference id
PurchaseOrderFormFieldValue StringPurchase order reference number
InvoiceIdFormFieldValue StringId of the invoice. Sometimes referred to as InvoiceNumber
InvoiceDateFormFieldValue DateTimeDate when invoice was issued
DueDateFormFieldValue DateTimeDate payment of invoice is due
VendorNameFormFieldValue StringName of vendor associated with invoice
VendorTaxIdFormFieldValue StringTaxpayer Id associated with the vendor
VendorAddressFormFieldValue StringVendor mailing address
VendorAddressRecipientFormFieldValue StringRecipient name associated with vendor address
CustomerTaxIdFormFieldValue StringTaxpayer Id associated with the customer
CustomerAddressFormFieldValue StringCustomer mailing address
CustomerAddressRecipientFormFieldValue StringRecipient name associated with customer address
BillingAddressFormFieldValue StringExplicit billing address for customer
BillingAddressRecipientFormFieldValue StringRecipient name associated with billing address
ShippingAddressFormFieldValue StringShipping address of customer specified in invoice
ShippingAddressRecipientFormFieldValue StringRecipient name associated with shipping address
ServiceAddressFormFieldValue StringExplicit service address of customer associated with invoice
ServiceAddressRecipientFormFieldValue StringRecipient name associated with service address
RemittanceAddressFormFieldValue StringExplicit remittance/payment address associated with customer
RemittanceAddressRecipientFormFieldValue StringRecipient name associated with remittance address
PaymentTermFormFieldValue StringTerms of payment specified in invoice
ServiceStartDateFormFieldValue DateTimeStarting date of provided service
ServiceEndDateFormFieldValue DateTimeEnd date of provided service
AmountDueFormFieldValue FloatTotal amount due
SubTotalFormFieldValue FloatSubtotal identified in invoice
PreviousUnpaidBalanceFormFieldValue FloatPrevious unpaid balance amount identified in invoice
InvoiceTotalFormFieldValue FloatTotal charges associated with the invoice
TotalTaxFormFieldValue FloatTotal tax amount identified in invoice
TotalVATFormFieldValue FloatTotal Value Added Tax amount identified in invoice
LineItemsInvoiceLineItemListing of invoice line items identified in invoice
FormTypeFormTypeType of detected form. Can be an Invoice, Receipt, BusinessCard, or IdentityCard
FormTypeConfidenceScoreFloatConfidence score in range 0 to 100 of the detected form type
PagesFormPageListing of pages identified in invoice
FieldsFormFieldDictionary of fields identified in the invoice. The form field name is used as a key in the dictionary

BusinessCard

Represents the model of a business card

Properties

PropertyTypeDescription
ContactNamesFormFieldValue List of StringContact Name(s)
FirstNameFormFieldValue StringFirst (given) name of contact.
LastNameFormFieldValue StringLast (family) name of contact
CompanyNamesFormFieldValue List of StringCompany Name(s) of contact
DepartmentNamesFormFieldValue List of StringDepartment Name(s)/Organization Name(s) of contact
JobTitlesFormFieldValue List of StringJob Title(s) of contact
EmailAddressesFormFieldValue List of StringEmail Address(es) of contact
WebsitesFormFieldValue List of StringWebsite(s) of contact
AddressesFormFieldValue List of StringAddress(es) of contact
MobilePhonesFormFieldValue List of StringMobile Phone Number(s) of contact
WorkPhonesFormFieldValue List of StringWork Phone Number(s) of contact
OtherPhonesFormFieldValue List of StringOther Phone Number(s) of contact
FaxesFormFieldValue List of StringFax Number(s) of contact
FormTypeFormTypeType of detected form. Can be an Invoice, Receipt, BusinessCard, or IdentityCard
FormTypeConfidenceScoreFloatConfidence score in range 0 to 100 of the detected form type
PagesFormPageListing of pages identified in business card
FieldsFormFieldDictionary of fields identified in the business card. The form field name is used as a key in the dictionary

IdentityDocument

Represents an identity document like a driver's license, passport, etc

Properties

PropertyTypeDescription
IdentityDocumentTypeFormFieldValue StringType of identity document. Possible values are a driver's license and passport.
FirstNameFormFieldValue StringFirst (given) name from document.
LastNameFormFieldValue StringLast (family) name from document
MiddleNameFormFieldValue StringMiddle name from document
SuffixFormFieldValue StringSuffix of name from document
DocumentNumberFormFieldValue StringIdentity document number
DateOfBirthFormFieldValue DateTimeDate of birth extracted from identity document
DateOfIssueFormFieldValue DateTimeIssue date extracted from identity document
DateOfExpirationFormFieldValue DateTimeExpiry date extracted from identity document
PlaceOfBirthFormFieldValue StringPlace of birth extracted from document
CountryCodeFormFieldValue StringISO 3166 country code
NationalityCodeFormFieldValue StringISO 3166 country code of nationality extracted from passport document
SexFormFieldValue StringPerson sex extracted from document. Possible values include M, F, and X
MachineReadableZoneFormFieldValue StringMRZ extracted from passport document. Two lines of 44 characters each
RegionFormFieldValue StringRegion extracted from identity document. (Driver's license)
AddressFormFieldValue StringPerson Address extracted from identity document. (Driver's license)
AddressCityFormFieldValue StringPerson Address city extracted from identity document. (Driver's license)
AddressStateFormFieldValue StringPerson Address state code extracted from identity document. (Driver's license)
AddressStateNameFormFieldValue StringPerson Address state name extracted from identity document. (Driver's license)
AddressZipCodeFormFieldValue StringPerson Address zip code extracted from identity document. (Driver's license)
AddressCountyFormFieldValue StringPerson Address county extracted from identity document. (Driver's license)
LicenseClassFormFieldValue StringLicense class extracted from identity document. (Driver's license)
RestrictionsFormFieldValue StringRestrictions extracted from identity document. (Driver's license)
EndorsementsFormFieldValue StringEndorsements extracted from identity document. (Driver's license)
VeteranFormFieldValue StringVeteran information extracted from identity document. (Driver's license)
FormTypeFormTypeType of detected form. Can be an Invoice, Receipt, BusinessCard, or IdentityCard
FormTypeConfidenceScoreFloatConfidence score in range 0 to 100 of the detected form type
PagesFormPageListing of pages identified in identity document
FieldsFormFieldDictionary of fields identified in identity document. The form field name is used as a key in the dictionary

ReceiptLineItem

Represents a line item inside a sales receipt

Properties

PropertyTypeDescription
ItemNameFormFieldValue StringItem name/Item description
QuantityFormFieldValue FloatQuantity detected in the line item
UnitPriceFormFieldValue FloatPrice per unit detected in the line item
TotalPriceFormFieldValue FloatTotal price of receipt line item

Receipt

Represents a sales receipt

Properties

PropertyTypeDescription
ReceiptTypeFormFieldValue StringType of sales receipt. e.g., itemized, hotel, etc.
MerchantNameFormFieldValue StringMerchant name identified in sales receipt
MerchantPhoneNumberFormFieldValue StringMerchant phone number identified in sales receipt
MerchantAddressFormFieldValue StringMerchant address identified in sales receipt
TransactionDateFormFieldValue DateTimeDate when the receipt was issued
TransactionTimeFormFieldValue TimeSpanTime when the receipt was issued
TotalFormFieldValue FloatTransaction total of the sales receipt
SubtotalFormFieldValue FloatSubtotal of the sales receipt (normally before tax is applied)
TaxFormFieldValue FloatTotal tax on receipt
TipFormFieldValue FloatTip indicated in the sales receipt
LineItemsReceiptLineItemListing of receipt line items identified in sales receipt
FormTypeFormTypeType of detected form. Can be an Invoice, Receipt, BusinessCard, or IdentityCard
FormTypeConfidenceScoreFloatConfidence score in range 0 to 100 of the detected form type
PagesFormPageListing of pages identified in receipt
FieldsFormFieldDictionary of fields identified in receipt. The form field name is used as a key in the dictionary

DetectedObject

It represents the information about the detected object by cognitive services.

Properties

PropertyTypeDescription
ObjectNameStringLanguage name.
ConfidenceScoreDoubleScore between 0 and 100 indicating the confidence for the detected object.

TranslatedText

It represents translated text information obtained from cognitive services.

Properties

PropertyTypeDescription
LanguageStringLanguage name.
LanguageCodeStringShortcode of language in ISO 639-1 format. Refer to List of ISO 639-1 codes
TextStringTranslated text.