Class DeepSORT<T>

Namespace

AiDotNet.ComputerVision.Tracking

Assembly

AiDotNet.dll

DeepSORT (Deep SORT) tracking with appearance features.

public class DeepSORT<T> : ObjectTrackerBase<T>

Type Parameters

T

The numeric type used for calculations.

Inheritance

object

ObjectTrackerBase<T>

DeepSORT<T>

Inherited Members

ObjectTrackerBase<T>.NumOps

ObjectTrackerBase<T>.Options

ObjectTrackerBase<T>.Tracks

ObjectTrackerBase<T>.NextTrackId

ObjectTrackerBase<T>.FrameCount

ObjectTrackerBase<T>.GetConfirmedTracks()

ObjectTrackerBase<T>.ComputeIoU(BoundingBox<T>, BoundingBox<T>)

ObjectTrackerBase<T>.ComputeCosineSimilarity(Tensor<T>, Tensor<T>)

ObjectTrackerBase<T>.HungarianAssignment(double[,])

ObjectTrackerBase<T>.CreateTrack(Detection<T>)

ObjectTrackerBase<T>.UpdateTrack(Track<T>, Detection<T>)

ObjectTrackerBase<T>.PredictNextPosition(Track<T>)

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

For Beginners: DeepSORT extends SORT by adding deep appearance features (Re-ID embeddings) to improve association accuracy, especially for occlusions and ID switches. It uses a cascade matching strategy.

Key features: - Deep appearance descriptor (Re-ID network) - Cascade matching (prioritize recent tracks) - Combined motion and appearance matching - Gallery of appearance features per track

Reference: Wojke et al., "Simple Online and Realtime Tracking with a Deep Association Metric", ICIP 2017

Constructors

DeepSORT(TrackingOptions<T>)

Creates a new DeepSORT tracker.

public DeepSORT(TrackingOptions<T> options)

Parameters

options TrackingOptions<T>

Properties

Name

Name of this tracker.

public override string Name { get; }

Property Value

string

Methods

Reset()

Resets the tracker state.

public override void Reset()

Update(List<Detection<T>>)

Updates tracks with new detections.

public override TrackingResult<T> Update(List<Detection<T>> detections)

Parameters

detections List<Detection<T>>

Detections from the current frame.

Returns

TrackingResult<T>

Tracking result with updated tracks.

Update(List<Detection<T>>, Tensor<T>?)

Updates tracks with new detections and image for appearance features.

public override TrackingResult<T> Update(List<Detection<T>> detections, Tensor<T>? image)

Parameters

detections List<Detection<T>>

Detections from the current frame.

image Tensor<T>

Current frame image.

Returns

TrackingResult<T>

Tracking result with updated tracks.