import { TrainingParameters as PoseTrainingParams } from './gtm-pose/teachable-posenet';
import { TeachablePoseNet, Metadata as PoseMetadata } from './gtm-pose';
import * as tf from '@tensorflow/tfjs';
import type { TeachableModel, ExplainedPredictionsOutput, TMType } from './TeachableModel';
interface TrainingParameters extends PoseTrainingParams {
}
interface BaseMetadata {
    modelBaseUrl?: string;
}
export declare type Metadata = BaseMetadata & PoseMetadata;
export default class PoseModel implements TeachableModel {
    protected model?: TeachablePoseNet;
    protected _ready?: Promise<boolean>;
    protected trained: boolean;
    protected busy: boolean;
    protected imageSize: number;
    protected _disposed: boolean;
    variant: TMType;
    explained?: HTMLCanvasElement;
    modelBaseUrl: string;
    private lastPose?;
    constructor(type: TMType, metadata?: Metadata, model?: tf.io.ModelJSON, weights?: ArrayBuffer);
    getVariant(): TMType;
    setXAICanvas(canvas: HTMLCanvasElement): void;
    protected load(metadata?: PoseMetadata, model?: tf.io.ModelJSON, weights?: ArrayBuffer): Promise<void>;
    ready(): Promise<boolean>;
    /**
     * If a pose is available, draw the keypoints and skeleton.
     *
     * @param image Image to draw the pose into.
     */
    draw(image: HTMLCanvasElement): HTMLCanvasElement;
    /**
     * Estimate pose if this is a PoseNet model, otherwise do nothing.
     * This caches the pose so draw() can use it without re-estimating.
     *
     * @param image Input image at correct resolution
     */
    estimate(image: HTMLCanvasElement): Promise<HTMLCanvasElement>;
    predict(image: HTMLCanvasElement): Promise<ExplainedPredictionsOutput>;
    /**
     * Predict directly from pose output data (for validation/internal use)
     */
    predictFromPoseData(poseData: Float32Array): Promise<ExplainedPredictionsOutput>;
    train(params: TrainingParameters, callbacks: tf.CustomCallbackArgs): Promise<tf.LayersModel>;
    addExample(className: number, image: HTMLCanvasElement): Promise<void>;
    dispose(): void;
    setName(name: string): void;
    getModel(): TeachablePoseNet | undefined;
    getImageSize(): number;
    isTrained(): boolean;
    isReady(): boolean;
    setSeed(seed: string): void;
    getMetadata(): PoseMetadata;
    save(handler: tf.io.IOHandler): Promise<tf.io.SaveResult>;
    setLabels(labels: string[]): void;
    getLabels(): string[];
    getLabel(ix: number): string;
    getNumExamples(): number;
    getExamplesPerClass(): number[];
    getNumValidation(): number;
    calculateAccuracy(): Promise<{
        reference: any;
        predictions: any;
    }>;
}
export {};