View Source TFLiteElixir.ObjectDetection (tflite_elixir v1.0.0)

Experimental object detection module.

Summary

Types

detection()

Functions

child_spec(init_arg)

Returns a specification to start this module under a supervisor.

predict(pid, input_path, opts \\ [])

Run the model against an image.

set_label(pid, label_file)

Give the detector its labels, either as a list or as the path to a file holding one label per line.

start(model, opts \\ [])

Start a detector for model, a path to a .tflite file or its contents.

Types

detection()

@type detection() :: %{
  class_id: integer(),
  score: float(),
  label: String.t() | nil,
  bbox: [integer()]
}

Functions

child_spec(init_arg)

Returns a specification to start this module under a supervisor.

See Supervisor.

predict(pid, input_path, opts \\ [])

@spec predict(
  pid(),
  binary()
  | %StbImage{data: term(), shape: term(), type: term()}
  | %Nx.Tensor{
      data: term(),
      names: term(),
      shape: term(),
      type: term(),
      vectorized_axes: term()
    },
  Keyword.t()
) :: [detection()]

Run the model against an image.

Options

:timeout - how long to wait for the answer, in milliseconds, or :infinity. Defaults to 30000. Raise it for a large model or a slow board; inference is local and bounded, so waiting is the right answer more often than giving up.

set_label(pid, label_file)

@spec set_label(pid(), String.t() | [String.t()]) :: :ok

Give the detector its labels, either as a list or as the path to a file holding one label per line.

start(model, opts \\ [])

@spec start(any(), any()) :: :ignore | {:error, any()} | {:ok, pid()}

Start a detector for model, a path to a .tflite file or its contents.

Options, all optional: :threshold (0.4) the score below which a detection is dropped, :labels (nil) a list of labels or the path to a file holding one per line, :jobs (System.schedulers_online/0) the interpreter's thread count, :use_tpu (false) and :tpu ("") to run on a named Edge TPU.