Object-Detector-App/p4ulor.obj.detector.data.domains.mediapipe

Package-level declarations

Types

Achievement

data class Achievement(val objectName: String, var certaintyScore: Float, var detectionDate: String? = null)

Model

enum Model(val id: String, @StringRes val description: Int) : Enum<Model>

These files should be in the assets folder with these names and file extension

MPImageResultListener

typealias MPImageResultListener = OutputHandler.ResultListener<ObjectDetectorResult, MPImage>

MyImageAnalyser

class MyImageAnalyser(val context: Context, val settings: ObjectDetectorSettings = ObjectDetectorSettings(), val resultsCallback: ObjectDetectorCallbacks) : ImageAnalysis.Analyzer

This class is the main class that uses the MediaPipe logic https://ai.google.dev/edge/mediapipe/solutions/vision/object_detector/android

ObjectDetectorCallbacks

interface ObjectDetectorCallbacks

Callbacks to expose callbacks set in MyImageAnalyser for MPImageResultListener, which is set at ObjectDetector.ObjectDetectorOptions.Builder.setResultListener. And for ErrorListener too.

ObjectDetectorSettings

data class ObjectDetectorSettings(val maxObjectDetections: Int = Companion.maxObjectDetections, val sensitivityThreshold: Float = 0.5f, val model: Model = Model.EFFICIENTDETV0, val processor: Delegate = Delegate.GPU, val mediaTypeToAnalyze: RunningMode = RunningMode.LIVE_STREAM)

A class that encapsulates the different settings that can be to detect objects using MediaPipe

ResultBundle

data class ResultBundle(val detectedObjects: ObjectDetectorResult, val processingTimeMs: Long, val inputImageHeight: Int, val inputImageWidth: Int)

Contains the data necessary to outline an object into the screen. Wraps results from inference, the time it takes for inference to be performed, and the input image and height for properly scaling UI to return back to callers

Properties

certaintyScore

val Detection.certaintyScore: String

Returns the

objectName

val Detection.objectName: String

The name of the detected object It's not clear why categories() is a list https://ai.google.dev/edge/api/mediapipe/python/mp/tasks/components/containers/Detection

Functions

calculatePoints

fun List<Achievement>.calculatePoints(): Float

getDonePercentage

fun List<Achievement>.getDonePercentage(): Float

pointsDifferenceBetween

fun List<Achievement>.pointsDifferenceBetween(other: List<UserAchievement>?): Float

Gets the diff between the local achievements this with the other in Firestore

reset

fun List<Achievement>.reset(): List<Achievement>

toUserAchievements

fun List<Achievement>.toUserAchievements(): List<UserAchievement>