Class FeatureMap<T>

Namespace

AiDotNet.ComputerVision.Detection.Backbones

Assembly

AiDotNet.dll

Feature map output from a backbone network.

public class FeatureMap<T>

Type Parameters

T

The numeric type used for calculations.

Inheritance

object

FeatureMap<T>

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Constructors

FeatureMap(Tensor<T>, int, int)

Creates a new feature map.

public FeatureMap(Tensor<T> features, int stride, int level)

Parameters

features Tensor<T>

The feature tensor.

stride int

The downsampling stride.

level int

The pyramid level index.

Properties

Channels

Gets the number of channels in the feature map.

public int Channels { get; }

Property Value

int

Features

The feature tensor with shape [batch, channels, height, width].

public Tensor<T> Features { get; set; }

Property Value

Tensor<T>

Level

The level index in the feature pyramid (0 = highest resolution).

public int Level { get; set; }

Property Value

int

SpatialSize

Gets the spatial dimensions of the feature map.

public (int Height, int Width) SpatialSize { get; }

Property Value

(int min, int max)

Stride

The stride (downsampling factor) of this feature map.

public int Stride { get; set; }

Property Value

int