vision Package

Azure AI Vision SDK for Python

Modules

diagnostics_logging

Classes that handle Vision SDK logging for troubleshooting purposes. If you report an issue to Microsoft, you may be asked to provide a Vision SDK log.

enums
image_analysis_client

Classes that represents the handling of Image Analysis functionalities.

image_analysis_data

Classes that represents the data structures of vision image analysis results.

interop
properties
version
vision_base_client

Classes that handles vision base functionalities for the various recognizers/analyzers.

Classes

ConsoleLogger

Represents a process-wide singleton that emits SDK log traces to the console.

FileLogger

Represents a process-wide singleton that emits SDK log traces to a provided file.

Frame

Represents a Frame being passed into or retrieved from the Vision SDK.

:param buffer:The memory that contains the frame.

FrameFormat

Represents a collection of image format properties (e.g. FOURCC, width, height, stride, ...)

FrameSource

Represents a source of image frame data, used as input to or output from Vision AI operations.

FrameSourceCallback

An interface that defines callback method used with FrameSource

Not fully implemented.

FrameWriter

Represents the ability to write image frame data, for use as input with Vision AI scenario operations.

ImageAnalysisErrorDetails

A representation of an error associated with an image analysis result.

ImageAnalysisEventArgs

Represents an asynchronous Image Analysis result payload as an event argument

ImageAnalysisOptions

Represents the configuration options that control the function of the ImageAnalyzer.

If you are doing Image Analysis using the standard model, you must set the features property to one or more visual features to analyze. There is no default selection for visual features. If you are using a custom model or doing Image Segmentation, you do not need to specify visual features.

ImageAnalysisResult

Represents the outcome of an Image Analysis operation.

Always start by checking the value of reason property to determine if the analysis operation was successful or not.

When an analysis operation is successful, applicable properties in this object will be populated based on the selected features (ImageAnalysisOptions.features) or custom-trained model (ImageAnalysisOptions.model_name). These results are parsed from the service JSON response. Other properties will be None.

Call ImageAnalysisResultDetails.from_result to get access to additional information about the result, such as the raw JSON response.

When analysis operation failed, call ImageAnalysisErrorDetails.from_result to get access to additional information about the error.

ImageAnalysisResultDetails

Represents additional information related to an image analysis result.

ImageAnalyzer

An object that facilitates Image Analysis operations with the Computer Vision service

ImageSourceBuffer

Represents a source of image data, used as input to or output from Vision AI operations.

Note

When used with ImageAnalyzer, callback is not supported. Call this constructor without arguments.

Then, using the image_writer property, get access to the ImageWriter object and write the image data to it.

ImageSourceBufferCallback

An interface that defines callback method used with ImageSourceBuffer.

Not implemented.

ImageWriter

Represents the ability to write image data, for use as input to Vision AI operations.

MemoryLogger

Represents a process-wide singleton that stores SDK log messages in a memory buffer.

VisionServiceOptions

Represents the configuration options used to connect to a remote Vision AI Service

key securely. For example, using Azure Key Vault. Do not compile the value of your key into your application source code. Fetch it at run-time when needed from a secure location.

VisionSource

Represents a source of vision data, used as input to a Computer Vision operation.