MediaPipeTasksVision Framework Reference

ObjectDetector

class ObjectDetector : NSObject

@brief Class that performs object detection on images.

The API expects a TFLite model with mandatory TFLite Model Metadata.

The API supports models with one image input tensor and one or more output tensors. To be more specific, here are the requirements:

Input tensor (kTfLiteUInt8/kTfLiteFloat32)

  • image input of size [batch x height x width x channels].
  • batch inference is not supported (batch is required to be 1).
  • only RGB inputs are supported (channels is required to be 3).
  • if type is kTfLiteFloat32, NormalizationOptions are required to be attached to the metadata for input normalization.

Output tensors must be the 4 outputs of a DetectionPostProcess op, i.e:(kTfLiteFloat32) (kTfLiteUInt8/kTfLiteFloat32)

  • locations tensor of size [num_results x 4], the inner array representing bounding boxes in the form [top, left, right, bottom].
  • BoundingBoxProperties are required to be attached to the metadata and must specify type=BOUNDARIES and coordinate_type=RATIO. (kTfLiteFloat32)
  • classes tensor of size [num_results], each value representing the integer index of a class.
  • optional (but recommended) label map(s) can be attached as AssociatedFiles with type TENSOR_VALUE_LABELS, containing one label per line. The first such AssociatedFile (if any) is used to fill the class_name field of the results. The display_name field is filled from the AssociatedFile (if any) whose locale matches the display_names_locale field of the ObjectDetectorOptions used at creation time (“en” by default, i.e. English). If none of these are available, only the index field of the results will be filled. (kTfLiteFloat32)
  • scores tensor of size [num_results], each value representing the score of the detected object.
  • optional score calibration can be attached using ScoreCalibrationOptions and an AssociatedFile with type TENSOR_AXIS_SCORE_CALIBRATION. See metadata_schema.fbs [1] for more details. (kTfLiteFloat32)
  • integer num_results as a tensor of size [1]

  • Creates a new instance of ObjectDetector from an absolute path to a TensorFlow Lite model file stored locally on the device and the default ObjectDetector.

    Declaration

    Swift

    convenience init(modelPath: String) throws

    Parameters

    modelPath

    An absolute path to a TensorFlow Lite model file stored locally on the device.
    error

    An optional error parameter populated when there is an error in initializing the object detector.

    Return Value

    A new instance of ObjectDetector with the given model path. nil if there is an error in initializing the object detector.

  • Creates a new instance of ObjectDetector from the given ObjectDetectorOptions.

    Declaration

    Swift

    init(options: ObjectDetectorOptions) throws

    Parameters

    options

    The options of type ObjectDetectorOptions to use for configuring the ObjectDetector.
    error

    An optional error parameter populated when there is an error in initializing the object detector.

    Return Value

    A new instance of ObjectDetector with the given options. nil if there is an error in initializing the object detector.

  • Performs object detection on the provided MPImage using the whole image as region of interest. Rotation will be applied according to the orientation property of the provided MPImage. Only use this method when the ObjectDetector is created with .image.

    This method supports detecting objects in RGBA images. If your MPImage has a source type of .pixelBuffer or .sampleBuffer, the underlying pixel buffer must use kCVPixelFormatType_32BGRA as its pixel format.

    If your MPImage has a source type of .image ensure that the color space is RGB with an Alpha channel.

    Declaration

    Swift

    func detect(image: MPImage) throws -> ObjectDetectorResult

    Parameters

    image

    The .image on which object detection is to be performed.

    Return Value

    An ObjectDetectorResult object that contains a list of detections, each detection has a bounding box that is expressed in the unrotated input frame of reference coordinates system, i.e. in [0,image_width) x [0,image_height), which are the dimensions of the underlying image data.

  • Performs object detection on the provided video frame of type MPImage using the whole image as region of interest. Rotation will be applied according to the orientation property of the provided MPImage. Only use this method when the ObjectDetector is created with .video.

    This method supports detecting objects in RGBA images. If your MPImage has a source type of .pixelBuffer or .sampleBuffer, the underlying pixel buffer must use kCVPixelFormatType_32BGRA as its pixel format.

    If your MPImage has a source type of .image ensure that the color space is RGB with an Alpha channel.

    Declaration

    Swift

    func detect(videoFrame image: MPImage, timestampInMilliseconds: Int) throws -> ObjectDetectorResult

    Parameters

    image

    The MPImage on which object detection is to be performed.
    timestampInMilliseconds

    The video frame’s timestamp (in milliseconds). The input timestamps must be monotonically increasing.

    Return Value

    An ObjectDetectorResult object that contains a list of detections, each detection has a bounding box that is expressed in the unrotated input frame of reference coordinates system, i.e. in [0,image_width) x [0,image_height), which are the dimensions of the underlying image data.

  • Sends live stream image data of type MPImage to perform object detection using the whole image as region of interest. Rotation will be applied according to the orientation property of the provided MPImage. Only use this method when the ObjectDetector is created with .liveStream.

    The object which needs to be continuously notified of the available results of object detection must confirm to ObjectDetectorLiveStreamDelegate protocol and implement the objectDetector(_:didFinishDetectionWithResult:timestampInMilliseconds:error:) delegate method.

    It’s required to provide a timestamp (in milliseconds) to indicate when the input image is sent to the object detector. The input timestamps must be monotonically increasing.

    This method supports detecting objects in RGBA images. If your MPImage has a source type of .pixelBuffer or .sampleBuffer, the underlying pixel buffer must use kCVPixelFormatType_32BGRA as its pixel format.

    If the input MPImage has a source type of .image ensure that the color space is RGB with an Alpha channel.

    If this method is used for detecting objects in live camera frames using AVFoundation, ensure that you request AVCaptureVideoDataOutput to output frames in kCMPixelFormat_32BGRA using its videoSettings property.

    Declaration

    Swift

    func detectAsync(image: MPImage, timestampInMilliseconds: Int) throws

    Parameters

    image

    A live stream image data of type MPImage on which object detection is to be performed.
    timestampInMilliseconds

    The timestamp (in milliseconds) which indicates when the input image is sent to the object detector. The input timestamps must be monotonically increasing.

    Return Value

    true if the image was sent to the task successfully, otherwise false.

  • Undocumented

  • Undocumented