<!--
{
  "availability" : [
    "iOS: 18.0.0 -",
    "iPadOS: 18.0.0 -",
    "macOS: 15.0.0 -",
    "tvOS: 18.0.0 -",
    "visionOS: 2.0.0 -",
    "watchOS: 27.0.0 -"
  ],
  "documentType" : "symbol",
  "framework" : "Vision",
  "identifier" : "/documentation/Vision/CoreMLFeatureValueObservation",
  "metadataVersion" : "0.1.0",
  "role" : "Structure",
  "symbol" : {
    "kind" : "Structure",
    "modules" : [
      "Vision"
    ],
    "preciseIdentifier" : "s:6Vision29CoreMLFeatureValueObservationV"
  },
  "title" : "CoreMLFeatureValueObservation"
}
-->

# CoreMLFeatureValueObservation

An object that represents a collection of key-value information that a Core ML image-analysis request produces.

```
struct CoreMLFeatureValueObservation
```

## Overview

This type of observation results from performing a [`CoreMLRequest`](/documentation/Vision/CoreMLRequest) image analysis with a <doc://com.apple.documentation/documentation/CoreML> model whose role is prediction rather than classification or image-to-image processing.

The framework infers that an <doc://com.apple.documentation/documentation/CoreML/MLModel> object is a predictor model if that model predicts multiple features. You can tell that a model predicts multiple features when its <doc://com.apple.documentation/documentation/CoreML/MLModel/modelDescription> object has a `nil` value for its <doc://com.apple.documentation/documentation/CoreML/MLModelDescription/predictedFeatureName> property, or when it inserts its output in an <doc://com.apple.documentation/documentation/CoreML/MLModelDescription/outputDescriptionsByName> dictionary.

The confidence for these observations is always `1.0`.

## Topics

### Creating an observation

[`init?(VNCoreMLFeatureValueObservation)`](/documentation/Vision/CoreMLFeatureValueObservation/init(_:))

Creates a feature value observation.

### Inspecting an observation

[`enum RequestDescriptor`](/documentation/Vision/RequestDescriptor)

A type that describes the request and revision combination.

### Getting the feature name and value

[`let featureName: String`](/documentation/Vision/CoreMLFeatureValueObservation/featureName)

The name in the model description of the model that produces this observation.

[`let featureValue: MLSendableFeatureValue`](/documentation/Vision/CoreMLFeatureValueObservation/featureValue)

The feature result of a request that outputs neither a classification nor an image.

## Relationships

### Conforms To

[`Copyable`](/documentation/Swift/Copyable)

[`Escapable`](/documentation/Swift/Escapable)

[`Hashable`](/documentation/Swift/Hashable)

[`Sendable`](/documentation/Swift/Sendable)

[`CustomStringConvertible`](/documentation/Swift/CustomStringConvertible)

[`VisionObservation`](/documentation/Vision/VisionObservation)

[`Equatable`](/documentation/Swift/Equatable)

[`Encodable`](/documentation/Swift/Encodable)

[`Decodable`](/documentation/Swift/Decodable)

[`SendableMetatype`](/documentation/Swift/SendableMetatype)

---

Copyright &copy; 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)