Details and Options

• ImageIdentify[{image₁,image₂,…},…] can be used to identify objects in multiple images.
• In ImageIdentify[image,category], possible forms for category include:
• "type"entity type, as used in Interpreter "concept"named concept, as used in "Concept" entities "word"English word, as used in WordData wordspecword sense specification, as used in WordData Entity[…]any appropriate entity category₁|category₂|…any of the category_i
• By default, ImageIdentify returns objects of the form Entity["Concept",…].
• The property "prop" can be one of the following:
• "Concept"a concept entity object "Entity"when possible, a concrete entity object "Probability"an association of concepts and probabilities "cprop"a property supported by "Concept" entities {prop₁,…}a list of property specifications
• The following options can be given:
• AcceptanceThreshold Automaticidentification acceptance threshold PerformanceGoal Automaticwhat to optimize in the identification SpecificityGoal Automaticwhat specificity of object type to seek TargetDevice"CPU"the target device on which to compute
• Possible settings for PerformanceGoal include "Speed" and "Quality".
• Possible settings for SpecificityGoal include:
• "Low"favor general categories of objects "High"favor specific kinds of objects sspecificity between 0 (lowest) and 1 (highest)
• When no content is found at an acceptable threshold, Missing["Unidentified"] is returned.
• ImageIdentify uses machine learning. Its methods, training sets and biases included therein may change and yield varied results in different versions of the Wolfram Language.
• ImageIdentify may download resources that will be stored in your local object store at $LocalBase, and can be listed using LocalObjects[] and removed using ResourceRemove.

Examples

Basic Examples  (2)

Identify the object present in the image:

Identify what type of dog is present in the image:

Scope  (3)

Return a result in a specific category:

Return a list of results:

Return a list of results and their associated probability:

Options  (4)

AcceptanceThreshold  (1)

The AcceptanceThreshold is selected automatically:

Specify a custom threshold:

If no identification is above the threshold, a Missing object is returned:

PerformanceGoal  (1)

Use PerformanceGoal"Speed" to get a result as fast as possible:

Use a slower, more accurate recognition:

SpecificityGoal  (2)

Privilege a result with a low specificity:

Privilege a result with a high specificity:

Use a custom value:

Get a table of identifications for different values of specificity:

Properties & Relations  (1)

The neural net used by ImageIdentify can be accessed using NetModel:

Possible Issues  (1)

If a recognition category is specified, probabilities are renormalized in that category:

Remove the category constraint to recognize the object correctly:

Neat Examples  (1)

Get 10 different identifications, along with their probabilities:

Visualize the result of object identification using a WordCloud: