Scrypted Documentation / ObjectDetection

Interface: ObjectDetection

ObjectDetection can run classifications or analysis on arbitrary media sources. E.g. TensorFlow, OpenCV, or a Coral TPU.

Methods

generateObjectDetections()

generateObjectDetections(videoFrames, session): Promise<AsyncGenerator<ObjectDetectionGeneratorResult, void, any>>

Parameters

• videoFrames: MediaObject | AsyncGenerator<VideoFrame, void, any>

• session: ObjectDetectionGeneratorSession

Returns

Promise<AsyncGenerator<ObjectDetectionGeneratorResult, void, any>>

---

detectObjects()

detectObjects(mediaObject, session?): Promise<ObjectsDetected>

Parameters

• mediaObject: MediaObject

• session?: ObjectDetectionSession

Returns

Promise<ObjectsDetected>

---

getDetectionModel()

getDetectionModel(settings?): Promise<ObjectDetectionModel>

Parameters

• settings?

Returns

Promise<ObjectDetectionModel>