The connection class for face landmarks.

Defines a rectangle, used e.g. as part of detection results or as input region-of-interest.

The connection class for pose landmarks.

This class wraps ImageFrame(CPU) & GpuBuffer(GPU) data. An instance of Image acts as an opaque reference to the underlying data objects. Image also ma

The connection class for hand landmarks.

Partitions the range [begin, end) into equal blocks of size grain_size each (except last one, might be less than grain_size).

TagMap allows access to a collection using a tag and index value. The underlying data in the collection is stored in a flat array. CollectionItemId is

Packet trace log event.

Helper class to compare semantic versions in terms of three integers, major, minor, and patch.

This class wraps a platform-specific buffer of GPU data. An instance of GpuBuffer acts as an opaque reference to the underlying data object.

A container for storing an image or a video frame. Formats supported by Image are listed in the ImageFormat enum. Pixels are encoded row-major in

Implementation of a unique file descriptor wrapper inspired by https://android.googlesource.com/platform/bionic/+/master/docs/fdsan.md This class is

* Parses the tasks.json configuration file. * * @property rootDir The root directory of the project, used to find tasks.json. */

RegistrationToken is a generic class that represents a registration that can be later undone, via a call to Unregister(). It is generally a good idea

A language code and its probability.

Partitions the range row_range x col_range into equal blocks of size row_range.grain_size() x col_range.grain_size() each (except last column and row

YuvBuffer represents a view over a YUV 4:2:0 image. YuvBuffers may be copied and moved efficiently; their backing buffers are shared and never deep c

Represents the classification results generated by {@link ImageClassifier}.

Represents the detection results generated by {@link FaceDetector}.

FloatBuffer represents a view over an interleaved floating-point image. FloatBuffers may be copied and moved efficiently; their backing buffers are s

A local registry of CalculatorGraphConfig definitions.

GrayBuffer represents a view over a grayscale (i.e. luminance, or Y-only) buffer. GrayBuffer may be copied and moved efficiently; their backing buffer

Lexicographic (first x, then y) comparator for MotionVector::pos.

Provides a shared ownership for a file descriptor. File descriptor is closed as soon as last SharedFd is destroyed. (Uses `std::shared_ptr` internall

A container for common tensor metadata information. Attributes: name: name of the tensor. description: description of what the tensor is.

Represents the classification results generated by {@link TextClassifier}.

A class which represents a timestamp in the calculator framework. There are several special values which can only be created with the static functions

RAII class for registration tokens: it calls Unregister() when it goes out of scope.

Calculator context manager owns and manages all calculator context objects of a calculator node.

RAII wrapper for EGL sync object.

Invoker for parallel execution. Thread storage is optional.

The face landmarks detection result from FaceLandmarker, where each vector element represents a single face detected in the image.

The hand landmarks detection result from HandLandmarker, where each vector element represents a single hand detected in the image.

Data are stored in column-major order by default.

The pose landmarks detection result from PoseLandmarker, where each vector element represents a single pose detected in the image.

Configuration for post-training quantization. Refer to https://www.tensorflow.org/lite/performance/post_training_quantization for different pos

This filter keeps track (on a window of specified size) of value changes over time, which as result gives us velocity of how value changes over time.

Utility class to allow for iterating over various containers of Tensors without* making any deep-copies or keeping any memory alive. Essentially this

Options to configure a thread. Default values are listed in the field descriptions.

Utility class that simplifies allowing (gating) multiple streams.

Audio format metadata. Attributes: num_channels: the number of channels of the audio data. sample_rate: the audio sample rate.

A container for the classification tensor metadata information. Attributes: label_files: information of the label files [1] in the classificati

Generates and loads a demo data set.

The DFS stack will contain a chain of nodes, from the root of the DFS to the current leaf.

Utility class that simplifies disallowing (gating) multiple streams.

The ctypes struct for FaceLandmarkerResult.