native library

Low-level generated bindings for the MediaPipe Tasks C API.

Most applications should use mp_audio, mp_text, or mp_vision. This library is public so those independently published packages can share one bundled native runtime.

The generated MpStatus FFI enum is hidden because package:mp_core.dart already exports the Dart-side MpStatus used throughout this SDK.

Classes

ImageEmbedderOptions: The options for configuring a MediaPipe image embedder task.
MpAudioClassifierInternal
MpAudioClassifierOptions: The options for configuring a MediaPipe AudioClassifier task.
MpAudioClassifierResult: The C representation of a list of audio classification results. The caller must call MpAudioClassifierCloseResult to free the memory.
MpAudioData: MediaPipe audio data.
MpBaseOptions: Base options for MediaPipe C Tasks.
MpCategories: A list of categories.
MpCategory: Defines a single classification result.
MpClassificationResult: Defines classification results of a model.
MpClassifications: Defines classification results for a given classifier head.
MpClassifierOptions: Classifier options for MediaPipe C classification Tasks.
MpDetection: Detection for a single bounding box.
MpDetectionResult: Detection results of a model.
MpEmbedderOptions: Embedder options for MediaPipe C embedding extraction tasks.
MpEmbedding: Embedding result for a given embedder head.
MpEmbeddingResult: Defines embedding results of a model.
MpFaceDetectorInternal
MpFaceDetectorOptions: The options for configuring a MediaPipe face detector task.
MpFaceLandmarkerInternal
MpFaceLandmarkerOptions: The options for configuring a MediaPipe face landmarker task.
MpFaceLandmarkerResult: The hand landmarker result from HandLandmarker, where each vector element represents a single hand detected in the image.
MpFlatbufferParserInternal
MpGestureRecognizerInternal
MpGestureRecognizerOptions: The options for configuring a MediaPipe gesture recognizer task.
MpGestureRecognizerResult: The gesture recognition result from GestureRecognizer, where each vector element represents a single hand detected in the image.
MpHandLandmarkerInternal
MpHandLandmarkerOptions: The options for configuring a MediaPipe hand landmarker task.
MpHandLandmarkerResult: The hand landmarker result from HandLandmarker, where each vector element represents a single hand detected in the image.
MpHolisticLandmarkerInternal
MpHolisticLandmarkerOptions: The options for configuring a MediaPipe holistic landmarker task.
MpHolisticLandmarkerResult: The holistic landmarks detection result from HolisticLandmarker.
MpImageClassifierInternal
MpImageClassifierOptions: The options for configuring a MediaPipe image classifier task.
MpImageEmbedderInternal
MpImageInternal
MpImageProcessingOptions: Options for image processing. If both region_of_interest and rotation are specified, the crop around the region-of-interest is extracted first, then the specified rotation is applied to the crop.
MpImageSegmenterInternal
MpImageSegmenterOptions: The options for configuring a mediapipe image segmenter task.
MpImageSegmenterResult
MpInteractiveSegmenterLegacyInternal
MpInteractiveSegmenterLegacyOptions: The options for configuring a mediapipe interactive segmenter task.
MpLandmark: Landmark represents a point in 3D space with x, y, z coordinates. The landmark coordinates are in meters. z represents the landmark depth, and the smaller the value the closer the world landmark is to the camera.
MpLandmarks: A list of MpLandmarks.
MpLanguageDetectorInternal
MpLanguageDetectorOptions: The options for configuring a MediaPipe language detector task.
MpLanguageDetectorPrediction: A language code and its probability.
MpLanguageDetectorResult: Task output.
MpMatrix: Data are stored in column-major order by default.
MpNormalizedKeypoint: A keypoint, defined by the coordinates (x, y), normalized by the image dimensions.
MpNormalizedLandmark: A normalized version of above MpLandmark struct. All coordinates should be within 0, 1.
MpNormalizedLandmarks: A list of MpNormalizedLandmarks.
MpObjectDetectorInternal
MpObjectDetectorOptions: The options for configuring a MediaPipe object detector task.
MpPoseLandmarkerInternal
MpPoseLandmarkerOptions: The options for configuring a MediaPipe pose landmarker task.
MpPoseLandmarkerResult: The hand landmarker result from PoseLandmarker, where each vector element represents a single hand detected in the image.
MpRect: Defines a rectangle, used e.g. as part of detection results or as input region-of-interest.
MpRectF: The coordinates are normalized wrt the image dimensions, i.e. generally in 0,1 but they may exceed these bounds if describing a region overlapping the image. The origin is on the top-left corner of the image.
MpRegionOfInterest: The Region-Of-Interest (ROI) to interact with.
MpSegmentationOptions: Options for configuring runtime behavior of ImageSegmenter.
MpStringList: Represents a list of strings.
MpTextClassifierInternal
MpTextClassifierOptions: The options for configuring a MediaPipe text classifier task.
MpTextEmbedderFormatContext: Encapsulates formatting instructions for models that require it (like Gecko).
MpTextEmbedderInternal
MpTextEmbedderOptions: The options for configuring a MediaPipe text embedder task.
NativeScope: Owns temporary native allocations for one task creation or invocation.
NativeTaskIsolate: A request/response worker that keeps one native task on one isolate.

Enums

MpAudioRunningMode: Supported processing modes.
MpDelegate: The delegate to run MediaPipe.
MpHostEnvironment: The environment that MediaPipe runs in.
MpHostSystem: Host OS on which MediaPipe tasks are running.
MpImageFormat: An enum describing supported raw image formats. Corresponds to mediapipe::ImageFormat::Format.
MpRegionOfInterestFormat: The format used to specify the region-of-interest.
MpRunningMode: Supported processing modes.
MpTextEmbedderEmbeddingType: The embedding type, used to format input text.
MpTextEmbedderRole: The role of the text in the context of the embedding task.

Functions

categoriesFromNative(MpCategories value) → List<Category>: Converts one native category list before its owner is released.
categoryFromNative(MpCategory value) → Category: Converts a native category.
classificationResultFromNative(MpClassificationResult value) → ClassificationResult: Converts a native classification result before its owner is released.
detectionResultFromNative(MpDetectionResult value, {int? timestampMs}) → DetectionResult: Converts a native detection result before its owner is released.
embeddingResultFromNative(MpEmbeddingResult value) → EmbeddingResult: Converts a native embedding result before its owner is released.
imageFromNative(MpImagePtr image, {String? task}) → MpImage: Copies a native image into an immutable Dart image.
landmarksFromNative(MpLandmarks value) → List<Landmark>: Converts one native world-landmark list before its owner is released.
matrixFromNative(MpMatrix value) → MpMatrix: Converts a native column-major matrix into the public row-major format.
MpAudioClassifierClassify(MpAudioClassifierPtr classifier, Pointer<MpAudioData> audio_data, Pointer<MpAudioClassifierResult> result_out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpAudioClassifierClassifyAsync(MpAudioClassifierPtr classifier, Pointer<MpAudioData> audio_data, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpAudioClassifierClose(MpAudioClassifierPtr classifier, Pointer<Pointer<Char>> error_msg) → MpStatus
MpAudioClassifierCloseResult(Pointer<MpAudioClassifierResult> result) → void: Frees the memory allocated inside a MpAudioClassifierResult result. Does not free the result pointer itself.
MpAudioClassifierCreate(Pointer<MpAudioClassifierOptions> options, Pointer<MpAudioClassifierPtr> classifier_out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpErrorFree(Pointer<Char> error_message) → void: Frees an error message.
MpFaceDetectorClose(MpFaceDetectorPtr detector, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFaceDetectorCloseResult(Pointer<MpFaceDetectorResult> result) → void: Frees the memory allocated inside a MpFaceDetectorResult result. Does not free the result pointer itself.
MpFaceDetectorCreate(Pointer<MpFaceDetectorOptions> options, Pointer<MpFaceDetectorPtr> detector, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFaceDetectorDetectAsync(MpFaceDetectorPtr detector, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFaceDetectorDetectForVideo(MpFaceDetectorPtr detector, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<MpFaceDetectorResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFaceDetectorDetectImage(MpFaceDetectorPtr detector, MpImagePtr image, Pointer<MpImageProcessingOptions> options, Pointer<MpFaceDetectorResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFaceLandmarkerClose(MpFaceLandmarkerPtr landmarker, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFaceLandmarkerCloseResult(Pointer<MpFaceLandmarkerResult> result) → void: Frees the memory allocated inside a MpFaceLandmarkerResult result. Does not free the result pointer itself.
MpFaceLandmarkerCreate(Pointer<MpFaceLandmarkerOptions> options, Pointer<MpFaceLandmarkerPtr> landmarker, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFaceLandmarkerDetectAsync(MpFaceLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFaceLandmarkerDetectForVideo(MpFaceLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<MpFaceLandmarkerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFaceLandmarkerDetectImage(MpFaceLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, Pointer<MpFaceLandmarkerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFlatbufferFreeString(Pointer<Char> str) → void: Frees a string allocated by MpFlatbufferGenerateText.
MpFlatbufferGenerateText(MpFlatbufferParser parser, Pointer<Uint8> buffer, int buffer_size, Pointer<Pointer<Char>> json_out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFlatbufferParserCreate(bool enable_strict_json, Pointer<MpFlatbufferParser> parser_out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpFlatbufferParserDelete(MpFlatbufferParser parser) → void: Deletes a Flatbuffer Parser.
MpFlatbufferParserGetError(MpFlatbufferParser parser) → Pointer<Char>: Gets the error message from the parser. The returned string is owned by the Parser and should not be freed.
MpFlatbufferParserParse(MpFlatbufferParser parser, Pointer<Char> source, Pointer<Pointer<Char>> error_msg) → MpStatus
MpGestureRecognizerClose(MpGestureRecognizerPtr recognizer, Pointer<Pointer<Char>> error_msg) → MpStatus
MpGestureRecognizerCloseResult(Pointer<MpGestureRecognizerResult> result) → void: Frees the memory allocated inside a MpGestureRecognizerResult result. Does not free the result pointer itself.
MpGestureRecognizerCreate(Pointer<MpGestureRecognizerOptions> options, Pointer<MpGestureRecognizerPtr> recognizer, Pointer<Pointer<Char>> error_msg) → MpStatus
MpGestureRecognizerRecognizeAsync(MpGestureRecognizerPtr recognizer, MpImagePtr image, Pointer<MpImageProcessingOptions> image_processing_options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpGestureRecognizerRecognizeForVideo(MpGestureRecognizerPtr recognizer, MpImagePtr image, Pointer<MpImageProcessingOptions> image_processing_options, int timestamp_ms, Pointer<MpGestureRecognizerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpGestureRecognizerRecognizeImage(MpGestureRecognizerPtr recognizer, MpImagePtr image, Pointer<MpImageProcessingOptions> image_processing_options, Pointer<MpGestureRecognizerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHandLandmarkerClose(MpHandLandmarkerPtr landmarker, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHandLandmarkerCloseResult(Pointer<MpHandLandmarkerResult> result) → void: Frees the memory allocated inside a MpHandLandmarkerResult result. Does not free the result pointer itself.
MpHandLandmarkerCreate(Pointer<MpHandLandmarkerOptions> options, Pointer<MpHandLandmarkerPtr> landmarker, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHandLandmarkerDetectAsync(MpHandLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHandLandmarkerDetectForVideo(MpHandLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<MpHandLandmarkerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHandLandmarkerDetectImage(MpHandLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, Pointer<MpHandLandmarkerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHolisticLandmarkerClose(MpHolisticLandmarkerPtr landmarker, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHolisticLandmarkerCloseResult(Pointer<MpHolisticLandmarkerResult> result) → void: Frees the memory allocated inside a MpHolisticLandmarkerResult result. Does not free the result pointer itself.
MpHolisticLandmarkerCreate(Pointer<MpHolisticLandmarkerOptions> options, Pointer<MpHolisticLandmarkerPtr> landmarker, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHolisticLandmarkerDetectAsync(MpHolisticLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHolisticLandmarkerDetectForVideo(MpHolisticLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<MpHolisticLandmarkerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpHolisticLandmarkerDetectImage(MpHolisticLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, Pointer<MpHolisticLandmarkerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageClassifierClassifyAsync(MpImageClassifierPtr classifier, MpImagePtr image, Pointer<MpImageProcessingOptions> image_processing_options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageClassifierClassifyForVideo(MpImageClassifierPtr classifier, MpImagePtr image, Pointer<MpImageProcessingOptions> image_processing_options, int timestamp_ms, Pointer<MpImageClassifierResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageClassifierClassifyImage(MpImageClassifierPtr classifier, MpImagePtr image, Pointer<MpImageProcessingOptions> image_processing_options, Pointer<MpImageClassifierResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageClassifierClose(MpImageClassifierPtr classifier, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageClassifierCloseResult(Pointer<MpImageClassifierResult> result) → void: Frees the memory allocated inside a MpImageClassifierResult result. Does not free the result pointer itself.
MpImageClassifierCreate(Pointer<MpImageClassifierOptions> options, Pointer<MpImageClassifierPtr> classifier, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageCreateFromFile(Pointer<Char> file_name, Pointer<MpImagePtr> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageCreateFromFloatData(MpImageFormat format, int width, int height, Pointer<Float> pixel_data, int pixel_data_size, Pointer<MpImagePtr> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageCreateFromImageFrame(MpImagePtr image, Pointer<MpImagePtr> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageCreateFromUint16Data(MpImageFormat format, int width, int height, Pointer<Uint16> pixel_data, int pixel_data_size, Pointer<MpImagePtr> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageCreateFromUint8Data(MpImageFormat format, int width, int height, Pointer<Uint8> pixel_data, int pixel_data_size, Pointer<MpImagePtr> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageDataFloat32(MpImagePtr image, Pointer<Pointer<Float>> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageDataUint16(MpImagePtr image, Pointer<Pointer<Uint16>> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageDataUint8(MpImagePtr image, Pointer<Pointer<Uint8>> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageEmbedderClose(MpImageEmbedderPtr embedder, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageEmbedderCloseResult(Pointer<ImageEmbedderResult> result) → void: Frees the memory allocated inside a ImageEmbedderResult result. Does not free the result pointer itself.
MpImageEmbedderCosineSimilarity(Pointer<MpEmbedding> u, Pointer<MpEmbedding> v, Pointer<Double> similarity, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageEmbedderCreate(Pointer<ImageEmbedderOptions> options, Pointer<MpImageEmbedderPtr> embedder, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageEmbedderEmbedAsync(MpImageEmbedderPtr embedder, MpImagePtr image, Pointer<MpImageProcessingOptions> image_processing_options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageEmbedderEmbedForVideo(MpImageEmbedderPtr embedder, MpImagePtr image, Pointer<MpImageProcessingOptions> image_processing_options, int timestamp_ms, Pointer<ImageEmbedderResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageEmbedderEmbedImage(MpImageEmbedderPtr embedder, MpImagePtr image, Pointer<MpImageProcessingOptions> image_processing_options, Pointer<ImageEmbedderResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageFree(MpImagePtr image) → void: Frees an MpImage.
MpImageGetByteDepth(MpImagePtr image) → int: Returns the byte depth of the image format. (e.g. 1 for SRGB, 2 for SRGB48).
MpImageGetChannels(MpImagePtr image) → int: Returns the number of channels in the image.
MpImageGetFormat(MpImagePtr image) → MpImageFormat
MpImageGetHeight(MpImagePtr image) → int: Returns the height of the image.
MpImageGetValueFloat32(MpImagePtr image, Pointer<Int> pos, int pos_size, Pointer<Float> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageGetValueUint16(MpImagePtr image, Pointer<Int> pos, int pos_size, Pointer<Uint16> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageGetValueUint8(MpImagePtr image, Pointer<Int> pos, int pos_size, Pointer<Uint8> out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageGetWidth(MpImagePtr image) → int: Returns the width of the image.
MpImageGetWidthStep(MpImagePtr image) → int: Returns the width step of the image.
MpImageIsAligned(MpImagePtr image, int alignment_boundary) → bool: Returns true if each row of the data is aligned to alignment_boundary.
MpImageIsContiguous(MpImagePtr image) → bool: Returns true if the pixel data is stored contiguously.
MpImageIsEmpty(MpImagePtr image) → bool: Returns true if the image is empty.
MpImageSegmenterClose(MpImageSegmenterPtr segmenter, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageSegmenterCloseResult(Pointer<MpImageSegmenterResult> result) → void: Frees the memory allocated inside a MpImageSegmenterResult result. Does not free the result pointer itself.
MpImageSegmenterCreate(Pointer<MpImageSegmenterOptions> options, Pointer<MpImageSegmenterPtr> segmenter, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageSegmenterGetLabels(MpImageSegmenterPtr segmenter, Pointer<MpStringList> label_list, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageSegmenterSegmentAsync(MpImageSegmenterPtr segmenter, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageSegmenterSegmentForVideo(MpImageSegmenterPtr segmenter, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<MpImageSegmenterResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageSegmenterSegmentImage(MpImageSegmenterPtr segmenter, MpImagePtr image, Pointer<MpImageProcessingOptions> options, Pointer<MpImageSegmenterResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpImageUsesGpu(MpImagePtr image) → bool: Returns true if the image is backed by GPU memory.
MpInteractiveSegmenterLegacyClose(MpInteractiveSegmenterLegacyPtr segmenter, Pointer<Pointer<Char>> error_msg) → MpStatus
MpInteractiveSegmenterLegacyCloseResult(Pointer<MpImageSegmenterResult> result) → void: Frees the memory allocated inside a MpImageSegmenterResult result. Does not free the result pointer itself.
MpInteractiveSegmenterLegacyCreate(Pointer<MpInteractiveSegmenterLegacyOptions> options, Pointer<MpInteractiveSegmenterLegacyPtr> segmenter, Pointer<Pointer<Char>> error_msg) → MpStatus
MpInteractiveSegmenterLegacySegmentImage(MpInteractiveSegmenterLegacyPtr segmenter, MpImagePtr image, Pointer<MpRegionOfInterest> roi, Pointer<MpImageProcessingOptions> image_processing_options, Pointer<MpImageSegmenterResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpLanguageDetectorClose(MpLanguageDetectorPtr detector, Pointer<Pointer<Char>> error_msg) → MpStatus
MpLanguageDetectorCloseResult(Pointer<MpLanguageDetectorResult> result) → void: Frees the memory allocated inside a MpLanguageDetectorResult result. Does not free the result pointer itself.
MpLanguageDetectorCreate(Pointer<MpLanguageDetectorOptions> options, Pointer<MpLanguageDetectorPtr> detector, Pointer<Pointer<Char>> error_msg) → MpStatus
MpLanguageDetectorDetect(MpLanguageDetectorPtr detector, Pointer<Char> utf8_str, Pointer<MpLanguageDetectorResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpObjectDetectorClose(MpObjectDetectorPtr detector, Pointer<Pointer<Char>> error_msg) → MpStatus
MpObjectDetectorCloseResult(Pointer<MpFaceDetectorResult> result) → void: Frees the memory allocated inside a MpObjectDetectorResult result. Does not free the result pointer itself.
MpObjectDetectorCreate(Pointer<MpObjectDetectorOptions> options, Pointer<MpObjectDetectorPtr> detector_out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpObjectDetectorDetectAsync(MpObjectDetectorPtr detector, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpObjectDetectorDetectForVideo(MpObjectDetectorPtr detector, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<MpFaceDetectorResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpObjectDetectorDetectImage(MpObjectDetectorPtr detector, MpImagePtr image, Pointer<MpImageProcessingOptions> options, Pointer<MpFaceDetectorResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpPoseLandmarkerClose(MpPoseLandmarkerPtr landmarker, Pointer<Pointer<Char>> error_msg) → MpStatus
MpPoseLandmarkerCloseResult(Pointer<MpPoseLandmarkerResult> result) → void: Frees the memory allocated inside a MpPoseLandmarkerResult result. Does not free the result pointer itself.
MpPoseLandmarkerCreate(Pointer<MpPoseLandmarkerOptions> options, Pointer<MpPoseLandmarkerPtr> landmarker_out, Pointer<Pointer<Char>> error_msg) → MpStatus
MpPoseLandmarkerDetectAsync(MpPoseLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<Pointer<Char>> error_msg) → MpStatus
MpPoseLandmarkerDetectForVideo(MpPoseLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, int timestamp_ms, Pointer<MpPoseLandmarkerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpPoseLandmarkerDetectImage(MpPoseLandmarkerPtr landmarker, MpImagePtr image, Pointer<MpImageProcessingOptions> options, Pointer<MpPoseLandmarkerResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpStringListFree(Pointer<MpStringList> string_list) → void: Frees an MpStringList.
MpTextClassifierClassify(MpTextClassifierPtr classifier, Pointer<Char> utf8_str, Pointer<MpImageClassifierResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
MpTextClassifierClose(MpTextClassifierPtr classifier, Pointer<Pointer<Char>> error_msg) → MpStatus
MpTextClassifierCloseResult(Pointer<MpImageClassifierResult> result) → void: Frees the memory allocated inside a MpTextClassifierResult result. Does not free the result pointer itself.
MpTextClassifierCreate(Pointer<MpTextClassifierOptions> options, Pointer<MpTextClassifierPtr> classifier, Pointer<Pointer<Char>> error_msg) → MpStatus
MpTextEmbedderClose(MpTextEmbedderPtr embedder, Pointer<Pointer<Char>> error_msg) → MpStatus
MpTextEmbedderCloseResult(Pointer<ImageEmbedderResult> result) → void: Frees the memory allocated inside a MpTextEmbedderResult result. Does not free the result pointer itself.
MpTextEmbedderCosSimilarity(Pointer<MpEmbedding> u, Pointer<MpEmbedding> v, Pointer<Double> similarity, Pointer<Pointer<Char>> error_msg) → MpStatus
MpTextEmbedderCreate(Pointer<MpTextEmbedderOptions> options, Pointer<MpTextEmbedderPtr> embedder, Pointer<Pointer<Char>> error_msg) → MpStatus
MpTextEmbedderEmbed(MpTextEmbedderPtr embedder, Pointer<Char> utf8_str, Pointer<MpTextEmbedderFormatContext> format_context, Pointer<ImageEmbedderResult> result, Pointer<Pointer<Char>> error_msg) → MpStatus
nativeString(Pointer<Char> value) → String?: Converts a nullable native UTF-8 pointer without taking ownership.
nativeTaskFailure(Object error) → Object: Rewrites an unresolvable native symbol into an actionable failure.
normalizedLandmarksFromNative(MpNormalizedLandmarks value) → List<NormalizedLandmark>: Converts one native normalized-landmark list before its owner is released.
resolveNativeModelAsset(ModelAsset asset) → Future<ModelAsset>: Resolves URI-backed model assets before crossing the native ABI.

Typedefs

ImageEmbedderResult = MpEmbeddingResult
MpAudioClassifierPtr = Pointer<MpAudioClassifierInternal>
MpFaceDetectorPtr = Pointer<MpFaceDetectorInternal>
MpFaceDetectorResult = MpDetectionResult
MpFaceLandmarkerPtr = Pointer<MpFaceLandmarkerInternal>
MpFlatbufferParser = Pointer<MpFlatbufferParserInternal>: Represents a Flatbuffer Parser.
MpGestureRecognizerPtr = Pointer<MpGestureRecognizerInternal>
MpHandLandmarkerPtr = Pointer<MpHandLandmarkerInternal>
MpHolisticLandmarkerPtr = Pointer<MpHolisticLandmarkerInternal>
MpImageClassifierPtr = Pointer<MpImageClassifierInternal>
MpImageClassifierResult = MpClassificationResult
MpImageEmbedderPtr = Pointer<MpImageEmbedderInternal>
MpImagePtr = Pointer<MpImageInternal>
MpImageSegmenterPtr = Pointer<MpImageSegmenterInternal>
MpInteractiveSegmenterLegacyPtr = Pointer<MpInteractiveSegmenterLegacyInternal>
MpLanguageDetectorPtr = Pointer<MpLanguageDetectorInternal>
MpObjectDetectorPtr = Pointer<MpObjectDetectorInternal>
MpObjectDetectorResult = MpDetectionResult
MpPoseLandmarkerPtr = Pointer<MpPoseLandmarkerInternal>
MpTextClassifierPtr = Pointer<MpTextClassifierInternal>
MpTextClassifierResult = MpClassificationResult
MpTextEmbedderPtr = Pointer<MpTextEmbedderInternal>
MpTextEmbedderResult = MpEmbeddingResult
NativeTaskWorkerFactory = FutureOr<NativeTaskWorkerHandler> Function(Object? initialMessage): Creates a command handler inside a dedicated native-task isolate.
NativeTaskWorkerHandler = FutureOr<Object?> Function(Object? command): Handles commands inside a dedicated native-task isolate.

native library

Classes

Enums

Functions

Typedefs

mp_core package

native library