ObjectDetectionModelBase

abstract class ObjectDetectionModelBase<I>(modelKindDescription: String?) : OnnxHighLevelModel<I, List<<ERROR CLASS>>>

Base class for object detection models.

Constructors

common

fun ObjectDetectionModelBase(modelKindDescription: String? = null)

Functions

convert

common

abstract fun convert(output: <ERROR CLASS>): List<<ERROR CLASS>>

Converts raw model output to the result.

detectObjects

common

fun detectObjects(image: I, topK: Int = 5): List<<ERROR CLASS>>

Returns the detected object for the given image sorted by the score.

initializeWith

common

open override fun initializeWith(vararg executionProviders: ExecutionProvider)

Initialize the model with the specified executions providers.

predict

common

open fun predict(input: I): List<<ERROR CLASS>>

Makes prediction on the given input.

summary

common

open fun summary(): <ERROR CLASS>

Properties

internalModel

common

abstract val internalModel: OnnxInferenceModel

Model used to make predictions.

modelKindDescription

common

open override val modelKindDescription: String? = null

High-level description of the model. Used for model summary printing. For the model hub models it equals to the string representation of the OnnxModelType

preprocessing

common

abstract val preprocessing: <ERROR CLASS><I, <ERROR CLASS><FloatArray, <ERROR CLASS>>>

Preprocessing operation specific to this model.

Inheritors

EfficientDetObjectDetectionModelBase

SSDLikeModelBase

Extensions

detectObjects

android

fun ObjectDetectionModelBase<Bitmap>.detectObjects(imageProxy: ImageProxy, topK: Int = 3): List<DetectedObject>

Returns the detected object for the given image sorted by the score. Internal preprocessing is updated to rotate image to match target orientation. After prediction, internal preprocessing is restored to the original state.