SSDObjectDetectionModel

class SSDObjectDetectionModel(internalModel: OnnxInferenceModel, modelKindDescription: String?) : SSDLikeModelBase<BufferedImage> , InferenceModel

Special model class for detection objects on images with built-in preprocessing and post-processing.

It internally uses ONNXModels.ObjectDetection.SSD trained on the COCO dataset.

Note that output class labels do not correspond to ids in COCO annotations. If you want to evaluate this model on the COCO validation/test set, you need to convert class predictions using appropriate mapping.

Since

0.3

Parameters

internalModel

model used to make predictions

Constructors

SSDObjectDetectionModel

jvm

fun SSDObjectDetectionModel(pathToModel: String)

Constructs the object detection model from a given path.

SSDObjectDetectionModel

jvm

fun SSDObjectDetectionModel(internalModel: OnnxInferenceModel, modelKindDescription: String? = null)

Functions

jvm

open override fun close()

convert

jvm

open override fun convert(output: OrtSession.Result): List<DetectedObject>

Converts raw model output to the result.

copy

jvm

open override fun copy(copiedModelName: String?, saveOptimizerState: Boolean, copyWeights: Boolean): SSDObjectDetectionModel

detectObjects

jvm

fun detectObjects(image: BufferedImage, topK: Int = 5): List<DetectedObject>

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

fun detectObjects(imageFile: File, topK: Int = 5): List<DetectedObject>

Returns the top N detected object for the given image file sorted by the score.

initializeWith

jvm

open override fun initializeWith(vararg executionProviders: ExecutionProvider)

Initialize the model with the specified executions providers.

predict

jvm

open override fun predict(inputData: FloatArray): Int

open fun predict(input: BufferedImage): List<DetectedObject>

Makes prediction on the given input.

predictSoftly

jvm

open override fun predictSoftly(inputData: FloatArray, predictionTensorName: String): FloatArray

reshape

jvm

open override fun reshape(vararg dims: Long)

summary

jvm

open override fun summary(): ModelSummary

Properties

inputDimensions

jvm

open override val inputDimensions: LongArray

internalModel

jvm

open override val internalModel: OnnxInferenceModel

modelKindDescription

jvm

open override val modelKindDescription: String?

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

jvm

open override val preprocessing: Operation<BufferedImage, Pair<FloatArray, TensorShape>>

Preprocessing operation specific to this model.

SSDObjectDetectionModel

Since

See also

Parameters

Constructors

Functions

Properties