# kornia.feature¶

## Detectors¶

Compute the Shi-Tomasi cornerness function.

Function does not do any normalization or nms. The response map is computed according the following formulation:

$R = min(eig(M))$

where:

$\begin{split}M = \sum_{(x,y) \in W} \begin{bmatrix} I^{2}_x & I_x I_y \\ I_x I_y & I^{2}_y \\ \end{bmatrix}\end{split}$
Parameters:
• input (Tensor) – input image with shape $$(B, C, H, W)$$.

• grads_mode (str, optional) – can be 'sobel' for standalone use or 'diff' for use on Gaussian pyramid. Default: "sobel"

• sigmas (, optional) – coefficients to be multiplied by multichannel response. Should be shape of $$(B)$$ It is necessary for performing non-maxima-suppression across different scale pyramid levels. See vlfeat. Default: None

Return type:

Tensor

Returns:

the response map per channel with shape $$(B, C, H, W)$$.

Example

>>> input = torch.tensor([[[
...    [0., 0., 0., 0., 0., 0., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 0., 0., 0., 0., 0., 0.],
... ]]])  # 1x1x7x7
>>> # compute the response map
gftt_response(input)
tensor([[[[0.0155, 0.0334, 0.0194, 0.0000, 0.0194, 0.0334, 0.0155],
[0.0334, 0.0575, 0.0339, 0.0000, 0.0339, 0.0575, 0.0334],
[0.0194, 0.0339, 0.0497, 0.0000, 0.0497, 0.0339, 0.0194],
[0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000],
[0.0194, 0.0339, 0.0497, 0.0000, 0.0497, 0.0339, 0.0194],
[0.0334, 0.0575, 0.0339, 0.0000, 0.0339, 0.0575, 0.0334],
[0.0155, 0.0334, 0.0194, 0.0000, 0.0194, 0.0334, 0.0155]]]])

kornia.feature.harris_response(input, k=0.04, grads_mode='sobel', sigmas=None)

Compute the Harris cornerness function.

Function does not do any normalization or nms. The response map is computed according the following formulation:

$R = max(0, det(M) - k \cdot trace(M)^2)$

where:

$\begin{split}M = \sum_{(x,y) \in W} \begin{bmatrix} I^{2}_x & I_x I_y \\ I_x I_y & I^{2}_y \\ \end{bmatrix}\end{split}$

and $$k$$ is an empirically determined constant $$k ∈ [ 0.04 , 0.06 ]$$

Parameters:
• input (Tensor) – input image with shape $$(B, C, H, W)$$.

• k (, optional) – the Harris detector free parameter. Default: 0.04

• grads_mode (str, optional) – can be 'sobel' for standalone use or 'diff' for use on Gaussian pyramid. Default: "sobel"

• sigmas (, optional) –

coefficients to be multiplied by multichannel response. Should be shape of $$(B)$$ It is necessary for performing non-maxima-suppression across different scale pyramid levels. See vlfeat. Default: None

Return type:

Tensor

Returns:

the response map per channel with shape $$(B, C, H, W)$$.

Example

>>> input = torch.tensor([[[
...    [0., 0., 0., 0., 0., 0., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 0., 0., 0., 0., 0., 0.],
... ]]])  # 1x1x7x7
>>> # compute the response map
harris_response(input, 0.04)
tensor([[[[0.0012, 0.0039, 0.0020, 0.0000, 0.0020, 0.0039, 0.0012],
[0.0039, 0.0065, 0.0040, 0.0000, 0.0040, 0.0065, 0.0039],
[0.0020, 0.0040, 0.0029, 0.0000, 0.0029, 0.0040, 0.0020],
[0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000],
[0.0020, 0.0040, 0.0029, 0.0000, 0.0029, 0.0040, 0.0020],
[0.0039, 0.0065, 0.0040, 0.0000, 0.0040, 0.0065, 0.0039],
[0.0012, 0.0039, 0.0020, 0.0000, 0.0020, 0.0039, 0.0012]]]])


Compute the absolute of determinant of the Hessian matrix.

Function does not do any normalization or nms. The response map is computed according the following formulation:

$R = det(H)$

where:

$\begin{split}M = \sum_{(x,y) \in W} \begin{bmatrix} I_{xx} & I_{xy} \\ I_{xy} & I_{yy} \\ \end{bmatrix}\end{split}$
Parameters:
• input (Tensor) – input image with shape $$(B, C, H, W)$$.

• grads_mode (str, optional) – can be 'sobel' for standalone use or 'diff' for use on Gaussian pyramid. Default: "sobel"

• sigmas (, optional) –

coefficients to be multiplied by multichannel response. Should be shape of $$(B)$$ It is necessary for performing non-maxima-suppression across different scale pyramid levels. See vlfeat. Default: None

Return type:

Tensor

Returns:

the response map per channel with shape $$(B, C, H, W)$$.

Shape:
• Input: $$(B, C, H, W)$$

• Output: $$(B, C, H, W)$$

Examples

>>> input = torch.tensor([[[
...    [0., 0., 0., 0., 0., 0., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 1., 1., 1., 1., 1., 0.],
...    [0., 0., 0., 0., 0., 0., 0.],
... ]]])  # 1x1x7x7
>>> # compute the response map
hessian_response(input)
tensor([[[[0.0155, 0.0334, 0.0194, 0.0000, 0.0194, 0.0334, 0.0155],
[0.0334, 0.0575, 0.0339, 0.0000, 0.0339, 0.0575, 0.0334],
[0.0194, 0.0339, 0.0497, 0.0000, 0.0497, 0.0339, 0.0194],
[0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000, 0.0000],
[0.0194, 0.0339, 0.0497, 0.0000, 0.0497, 0.0339, 0.0194],
[0.0334, 0.0575, 0.0339, 0.0000, 0.0339, 0.0575, 0.0334],
[0.0155, 0.0334, 0.0194, 0.0000, 0.0194, 0.0334, 0.0155]]]])

kornia.feature.dog_response(input)

Compute the Difference-of-Gaussian response.

Parameters:

input (Tensor) – a given the gaussian 5d tensor $$(B, C, D, H, W)$$.

Return type:

Tensor

Returns:

the response map per channel with shape $$(B, C, D-1, H, W)$$.

kornia.feature.dog_response_single(input, sigma1=1.0, sigma2=1.6)

Compute the Difference-of-Gaussian response.

Parameters:
Return type:

Tensor

Returns:

the response map per channel with shape $$(B, C, H, W)$$.

Module that calculates Hessian blobs.

See hessian_response() for details.

Module that calculates Shi-Tomasi corners.

See gftt_response() for details.

Module that calculates Harris corners.

See harris_response() for details.

class kornia.feature.BlobDoG

Module that calculates Difference-of-Gaussians blobs.

See :func: ~kornia.feature.dog_response for details.

class kornia.feature.BlobDoGSingle(sigma1=1.0, sigma2=1.6)

Module that calculates Difference-of-Gaussians blobs.

See dog_response_single() for details.

class kornia.feature.KeyNet(pretrained=False, keynet_conf=keynet_default_config)

Key.Net model definition – local feature detector (response function). This is based on the original code from paper “Key.Net: Keypoint Detection by Handcrafted and Learned CNN Filters”. See [BLRPM19] for more details.

Parameters:
• pretrained (bool, optional) – Download and set pretrained weights to the model. Default: False

• keynet_conf (KeyNet_conf, optional) – Dict with initialization parameters. Do not pass it, unless you know what you are doing. Default: keynet_default_config

Returns:

KeyNet response score.

Shape:
• Input: $$(B, 1, H, W)$$

• Output: $$(B, 1, H, W)$$

class kornia.feature.MultiResolutionDetector(model, num_features=2048, config=get_default_detector_config(), ori_module=None, aff_module=None)

Multi-scale feature detector, based on code from KeyNet. Can be used with any response function.

This is based on the original code from paper “Key.Net: Keypoint Detection by Handcrafted and Learned CNN Filters”. See [BLRPM19] for more details.

Parameters:

Three stage local feature detection. First the location and scale of interest points are determined by detect function. Then affine shape and orientation.

Parameters:

img (Tensor) – image to extract features with shape [1xCxHxW]. KeyNetDetector does not support batch processing,

because the number of detections is different on each image.
mask: a mask with weights where to apply the response function. The shape must be the same as

the input image.

Returns:

shape [1xNx2x3]. Detected local affine frames. responses: shape [1xNx1]. Response function values for corresponding lafs

Return type:

lafs

remove_borders(score_map, borders=15)

It removes the borders of the image to avoid detections on the corners.

Return type:

Tensor

class kornia.feature.ScaleSpaceDetector(num_features=500, mr_size=6.0, scale_pyr_module=ScalePyramid(3, 1.6, 15), resp_module=BlobHessian(), nms_module=ConvSoftArgmax3d((3, 3, 3), (1, 1, 1), (1, 1, 1), normalized_coordinates=False, output_value=True), ori_module=PassLAF(), aff_module=PassLAF(), minima_are_also_good=False, scale_space_response=False)

Module for differentiable local feature detection, as close as possible to classical local feature detectors like Harris, Hessian-Affine or SIFT (DoG).

It has 5 modules inside: scale pyramid generator, response (“cornerness”) function, soft nms function, affine shape estimator and patch orientation estimator. Each of those modules could be replaced with learned custom one, as long, as they respect output shape.

Parameters:
• num_features (int, optional) – Number of features to detect. In order to keep everything batchable, output would always have num_features output, even for completely homogeneous images. Default: 500

• mr_size (float, optional) – multiplier for local feature scale compared to the detection scale. 6.0 is matching OpenCV 12.0 convention for SIFT. Default: 6.0

• scale_pyr_module (Module, optional) – generates scale pyramid. See ScalePyramid for details. Default: ScalePyramid(3, 1.6, 10).

• resp_module (Module, optional) – calculates 'cornerness' of the pixel. Default: BlobHessian()

• nms_module (Module, optional) – outputs per-patch coordinates of the response maxima. See ConvSoftArgmax3d for details. Default: ConvSoftArgmax3d((3, 3, 3), (1, 1, 1), (1, 1, 1), normalized_coordinates=False, output_value=True)

• ori_module (Module, optional) – for local feature orientation estimation. Default:class:~kornia.feature.PassLAF, which does nothing. See LAFOrienter for details. Default: PassLAF()

• aff_module (Module, optional) – for local feature affine shape estimation. Default: PassLAF, which does nothing. See LAFAffineShapeEstimator for details.

• minima_are_also_good (bool, optional) – if True, then both response function minima and maxima are detected Useful for symmetric response functions like DoG or Hessian. Default is False Default: False

Three stage local feature detection. First the location and scale of interest points are determined by detect function. Then affine shape and orientation.

Parameters:
• img (Tensor) – image to extract features with shape [BxCxHxW]

• mask (, optional) – a mask with weights where to apply the response function. The shape must be the same as the input image. Default: None

Returns:

shape [BxNx2x3]. Detected local affine frames. responses: shape [BxNx1]. Response function values for corresponding lafs

Return type:

lafs

class kornia.feature.KeyNetDetector(pretrained=False, num_features=2048, keynet_conf=keynet_default_config, ori_module=None, aff_module=None)

Multi-scale feature detector based on KeyNet.

This is based on the original code from paper “Key.Net: Keypoint Detection by Handcrafted and Learned CNN Filters”. See [BLRPM19] for more details.

Parameters:
• pretrained (bool, optional) – Download and set pretrained weights to the model. Default: False

• num_features (int, optional) – Number of features to detect. Default: 2048

• keynet_conf (KeyNet_conf, optional) – Dict with initialization parameters. Do not pass it, unless you know what you are doing. Default: keynet_default_config

• ori_module (, optional) – for local feature orientation estimation. Default: PassLAF, which does nothing. See LAFOrienter for details.

• aff_module (, optional) – for local feature affine shape estimation. Default: PassLAF, which does nothing. See LAFAffineShapeEstimator for details.

Three stage local feature detection. First the location and scale of interest points are determined by detect function. Then affine shape and orientation.

Parameters:

img (Tensor) – image to extract features with shape [1xCxHxW]. KeyNetDetector does not support batch processing,

because the number of detections is different on each image.
mask: a mask with weights where to apply the response function. The shape must be the same as

the input image.

Returns:

shape [1xNx2x3]. Detected local affine frames. responses: shape [1xNx1]. Response function values for corresponding lafs

Return type:

lafs

## Descriptors¶

class kornia.feature.DenseSIFTDescriptor(num_ang_bins=8, num_spatial_bins=4, spatial_bin_size=4, rootsift=True, clipval=0.2, stride=1, padding=1)

Module, which computes SIFT descriptor densely over the image.

Parameters:
• num_ang_bins (int, optional) – Number of angular bins. (8 is default) Default: 8

• num_spatial_bins (int, optional) – Number of spatial bins per descriptor (4 is default). Default: 4

You might want to set odd number and relevant padding to keep feature map size

spatial_bin_size: Size of a spatial bin in pixels (4 is default) clipval: clipping value to reduce single-bin dominance rootsift: (bool) if True, RootSIFT (Arandjelović et. al, 2012) is computed stride: default 1 padding: default 0

Returns:

DenseSIFT descriptor of the image

Return type:

Tensor

Shape:
• Input: (B, 1, H, W)

• Output: (B, num_ang_bins * num_spatial_bins ** 2, (H+padding)/stride, (W+padding)/stride)

Examples::
>>> input =  torch.rand(2, 1, 200, 300)
>>> SIFT = DenseSIFTDescriptor()
>>> descs = SIFT(input) # 2x128x194x294

class kornia.feature.SIFTDescriptor(patch_size=41, num_ang_bins=8, num_spatial_bins=4, rootsift=True, clipval=0.2)

Module which computes SIFT descriptors of given patches.

Parameters:
• patch_size (int, optional) – Input patch size in pixels. Default: 41

• num_ang_bins (int, optional) – Number of angular bins. Default: 8

• num_spatial_bins (int, optional) – Number of spatial bins. Default: 4

• clipval (float, optional) – clipping value to reduce single-bin dominance Default: 0.2

• rootsift (bool, optional) – if True, RootSIFT (Arandjelović et. al, 2012) is computed. Default: True

Returns:

SIFT descriptor of the patches with shape.

Shape:
• Input: $$(B, 1, \text{num_spatial_bins}, \text{num_spatial_bins})$$

• Output: $$(B, \text{num_ang_bins * num_spatial_bins ** 2})$$

Example

>>> input = torch.rand(23, 1, 32, 32)
>>> SIFT = SIFTDescriptor(32, 8, 4)
>>> descs = SIFT(input) # 23x128

class kornia.feature.MKDDescriptor(patch_size=32, kernel_type='concat', whitening='pcawt', training_set='liberty', output_dims=128)

Module that computes Multiple Kernel local descriptors.

This is based on the paper “Understanding and Improving Kernel Local Descriptors”. See [MTB+19] for more details.

Parameters:
• patch_size (int, optional) – Input patch size in pixels. Default: 32

• kernel_type (str, optional) – Parametrization of kernel 'concat', 'cart', 'polar'. Default: "concat"

• whitening (str, optional) – Whitening transform to apply None, 'lw', 'pca', 'pcawt', 'pcaws'. Default: "pcawt"

• training_set (str, optional) – Set that model was trained on 'liberty', 'notredame', 'yosemite'. Default: "liberty"

• output_dims (int, optional) – Dimensionality reduction. Default: 128

Returns:

Explicit cartesian or polar embedding.

Shape:
• Input: $$(B, in_{dims}, fmap_{size}, fmap_{size})$$.

• Output: $$(B, out_{dims}, fmap_{size}, fmap_{size})$$,

Examples

>>> patches = torch.rand(23, 1, 32, 32)
>>> mkd = MKDDescriptor(patch_size=32,
...                     kernel_type='concat',
...                     whitening='pcawt',
...                     training_set='liberty',
...                     output_dims=128)
>>> desc = mkd(patches) # 23x128

class kornia.feature.HardNet(pretrained=False)

Module, which computes HardNet descriptors of given grayscale patches of 32x32.

This is based on the original code from paper “Working hard to know your neighbor’s margins: Local descriptor learning loss”. See [MMRM17] for more details.

Parameters:

pretrained (bool, optional) – Download and set pretrained weights to the model. Default: False

Returns:

HardNet descriptor of the patches.

Return type:

torch.Tensor

Shape:
• Input: $$(B, 1, 32, 32)$$

• Output: $$(B, 128)$$

Examples

>>> input = torch.rand(16, 1, 32, 32)
>>> hardnet = HardNet()
>>> descs = hardnet(input) # 16x128

class kornia.feature.HardNet8(pretrained=False)

Module, which computes HardNet8 descriptors of given grayscale patches of 32x32.

This is based on the original code from paper “Improving the HardNet Descriptor”. See [Pul20] for more details.

Parameters:

pretrained (bool, optional) – Download and set pretrained weights to the model. Default: False

Returns:

HardNet8 descriptor of the patches.

Return type:

torch.Tensor

Shape:
• Input: $$(B, 1, 32, 32)$$

• Output: $$(B, 128)$$

Examples

>>> input = torch.rand(16, 1, 32, 32)
>>> hardnet = HardNet8()
>>> descs = hardnet(input) # 16x128

class kornia.feature.HyNet(pretrained=False, is_bias=True, is_bias_FRN=True, dim_desc=128, drop_rate=0.3, eps_l2_norm=1e-10)

Module, which computes HyNet descriptors of given grayscale patches of 32x32.

This is based on the original code from paper “HyNet: Learning Local Descriptor with Hybrid Similarity Measure and Triplet Loss”. See [TBLN+20] for more details.

Parameters:
• pretrained (bool, optional) – Download and set pretrained weights to the model. Default: False

• is_bias (bool, optional) – use bias in TLU layers Default: True

• is_bias_FRN (bool, optional) – use bias in FRN layers Default: True

• dim_desc (int, optional) – descriptor dimensionality, Default: 128

• drop_rate (float, optional) – dropout rate, Default: 0.3

• eps_l2_norm (float, optional) – to avoid div by zero Default: 1e-10

Returns:

HyNet descriptor of the patches.

Shape:
• Input: $$(B, 1, 32, 32)$$

• Output: $$(B, 128)$$

Examples

>>> input = torch.rand(16, 1, 32, 32)
>>> hynet = HyNet()
>>> descs = hynet(input) # 16x128

class kornia.feature.TFeat(pretrained=False)

Module, which computes TFeat descriptors of given grayscale patches of 32x32.

This is based on the original code from paper “Learning local feature descriptors with triplets and shallow convolutional neural networks”. See [BRPM16] for more details

Parameters:

pretrained (bool, optional) – Download and set pretrained weights to the model. Default: False

Returns:

TFeat descriptor of the patches.

Return type:

torch.Tensor

Shape:
• Input: $$(B, 1, 32, 32)$$

• Output: $$(B, 128)$$

Examples

>>> input = torch.rand(16, 1, 32, 32)
>>> tfeat = TFeat()
>>> descs = tfeat(input) # 16x128

class kornia.feature.SOSNet(pretrained=False)

128-dimensional SOSNet model definition for 32x32 patches.

This is based on the original code from paper “SOSNet:Second Order Similarity Regularization for Local Descriptor Learning”.

Parameters:

pretrained (bool, optional) – Download and set pretrained weights to the model. Default: False

Shape:
• Input: $$(B, 1, 32, 32)$$

• Output: $$(B, 128)$$

Examples

>>> input = torch.rand(8, 1, 32, 32)
>>> sosnet = SOSNet()
>>> descs = sosnet(input) # 8x128

class kornia.feature.LAFDescriptor(patch_descriptor_module=None, patch_size=32, grayscale_descriptor=True)

Module to get local descriptors, corresponding to LAFs (keypoints).

Internally uses get_laf_descriptors().

Parameters:
forward(img, lafs)

Three stage local feature detection.

First the location and scale of interest points are determined by detect function. Then affine shape and orientation.

Parameters:
Return type:

Tensor

Returns:

Local descriptors of shape $$(B,N,D)$$ where $$D$$ is descriptor size.

class kornia.feature.SOLD2(pretrained=True, config=None)

Module, which detects and describe line segments in an image.

This is based on the original code from the paper “SOLD²: Self-supervised Occlusion-aware Line Detector and Descriptor”. See for more details.

Parameters:
• config (Optional[DetectorCfg], optional) – Dict specifying parameters. None will load the default parameters, which are tuned for images in the range 400~800 px. Default: None

• pretrained (bool, optional) – If True, download and set pretrained weights to the model. Default: True

Returns:

The raw junction and line heatmaps, the semi-dense descriptor map, as well as the list of detected line segments (ij coordinates convention).

Example

>>> images = torch.rand(2, 1, 512, 512)
>>> sold2 = SOLD2()
>>> outputs = sold2(images)
>>> line_seg1 = outputs["line_segments"][0]
>>> line_seg2 = outputs["line_segments"][1]
>>> desc1 = outputs["dense_desc"][0]
>>> desc2 = outputs["dense_desc"][1]
>>> matches = sold2.match(line_seg1, line_seg2, desc1[None], desc2[None])

forward(img)
Parameters:

img (Tensor) – batched images with shape $$(B, 1, H, W)$$.

Returns:

list of N line segments in each of the B images $$List[(N, 2, 2)]$$. - junction_heatmap: raw junction heatmap of shape $$(B, H, W)$$. - line_heatmap: raw line heatmap of shape $$(B, H, W)$$. - dense_desc: the semi-dense descriptor map of shape $$(B, 128, H/4, W/4)$$.

Return type:

• line_segments

kornia.feature.get_laf_descriptors(img, lafs, patch_descriptor, patch_size=32, grayscale_descriptor=True)

Function to get local descriptors, corresponding to LAFs (keypoints).

Parameters:
Return type:

Tensor

Returns:

Local descriptors of shape $$(B,N,D)$$ where $$D$$ is descriptor size.

## Local Features (Detector and Descriptors together)¶

class kornia.feature.LocalFeature(detector, descriptor, scaling_coef=1.0)

Module, which combines local feature detector and descriptor.

Parameters:
Parameters:
• img (Tensor) – image to extract features with shape $$(B,C,H,W)$$.

• mask (, optional) – a mask with weights where to apply the response function. The shape must be the same as the input image. Default: None

Return type:

Returns:

• Detected local affine frames with shape $$(B,N,2,3)$$.

• Response function values for corresponding lafs with shape $$(B,N,1)$$.

• Local descriptors of shape $$(B,N,D)$$ where $$D$$ is descriptor size.

class kornia.feature.SOLD2_detector(pretrained=True, config=None)

Module, which detects line segments in an image.

This is based on the original code from the paper “SOLD²: Self-supervised Occlusion-aware Line Detector and Descriptor”. See for more details.

Parameters:
• config (DetectorCfg) – Configuration object containing all parameters. None will load the default parameters, which are tuned for images in the range 400~800 px. Using a dataclass ensures type safety and clearer parameter management. Default: None

• pretrained (bool) – If True, download and set pretrained weights to the model. Default: True

Returns:

The raw junction and line heatmaps, as well as the list of detected line segments (ij coordinates convention).

Example

>>> img = torch.rand(1, 1, 512, 512)
>>> sold2_detector = SOLD2_detector()
>>> line_segments = sold2_detector(img)["line_segments"]

forward(img)
Parameters:

img (Tensor) – batched images with shape $$(B, 1, H, W)$$.

Returns:

list of N line segments in each of the B images $$List[(N, 2, 2)]$$. - junction_heatmap: raw junction heatmap of shape $$(B, H, W)$$. - line_heatmap: raw line heatmap of shape $$(B, H, W)$$.

Return type:

• line_segments

class kornia.feature.DeDoDe(detector_model='L', descriptor_model='G', amp_dtype=torch.float16)

Module which detects and/or describes local features in an image using the DeDode method.

See [EBWF24] for details.

Note

DeDode takes ImageNet normalized images as input (not in range [0, 1]).

Parameters:
• detector_model (Literal['L'], optional) – The detector model kind. Available options are: L. Default: "L"

• descriptor_model (Literal['G', 'B'], optional) – The descriptor model kind. Available options are: G or B Default: "G"

• amp_dtype (dtype, optional) – The automatic mixed precision desired. Default: torch.float16

Example

>>> dedode = DeDoDe.from_pretrained(detector_weights="L-C4-v2", descriptor_weights="B-upright")
>>> images = torch.randn(1, 3, 256, 256)
>>> keypoints, scores = dedode.detect(images)
>>> descriptions = dedode.describe(images, keypoints = keypoints)
>>> keypoints, scores, features = dedode(images) # alternatively do both

describe(images, keypoints=None, apply_imagenet_normalization=True)

Describes keypoints in the input images. If keypoints are not provided, returns the dense descriptors.

Parameters:
• images (Tensor) – A tensor of shape $$(B, 3, H, W)$$ containing the input images.

• keypoints (, optional) – An optional tensor of shape $$(B, N, 2)$$ containing the detected keypoints. Default: None

• apply_imagenet_normalization (bool, optional) – Whether to apply ImageNet normalization to the input images. Default: True

Returns:

A tensor of shape $$(B, N, DIM)$$ containing the descriptions of the detected keypoints. If the dense descriptors are requested, the shape is $$(B, DIM, H, W)$$.

Return type:

descriptions

detect(images, n=10000, apply_imagenet_normalization=True, pad_if_not_divisible=True, crop_h=None, crop_w=None)

Detects keypoints in the input images.

Parameters:
• images (Tensor) – A tensor of shape $$(B, 3, H, W)$$ containing the input images.

• n (, optional) – The number of keypoints to detect. Default: 10000

• apply_imagenet_normalization (bool, optional) – Whether to apply ImageNet normalization to the input images. Default: True

• crop_h (, optional) – The height of the crop to be used for detection. If None, the full image is used. Default: None

• crop_w (, optional) – The width of the crop to be used for detection. If None, the full image is used. Default: None

Returns:

A tensor of shape $$(B, N, 2)$$ containing the detected keypoints, normalized to the range $$[-1, 1]$$. scores: A tensor of shape $$(B, N)$$ containing the scores of the detected keypoints.

Return type:

keypoints

forward(images, n=10_000, apply_imagenet_normalization=True, pad_if_not_divisible=True)

Detects and describes keypoints in the input images.

Parameters:
• images (Tensor) – A tensor of shape $$(B, 3, H, W)$$ containing the ImageNet-Normalized input images.

• n (, optional) – The number of keypoints to detect. Default: 10_000

• apply_imagenet_normalization (bool, optional) – Whether to apply ImageNet normalization to the input images. Default: True

Returns:

A tensor of shape $$(B, N, 2)$$ containing the detected keypoints in the image range, unlike .detect() function scores: A tensor of shape $$(B, N)$$ containing the scores of the detected keypoints. descriptions: A tensor of shape $$(B, N, DIM)$$ containing the descriptions of the detected keypoints. DIM is 256 for B and 512 for G.

Return type:

keypoints

classmethod from_pretrained(detector_weights='L-C4-v2', descriptor_weights='G-upright', amp_dtype=torch.float16)

Loads a pretrained model.

Parameters:
Return type:

Module

Returns:

The pretrained model.

class kornia.feature.DISK(desc_dim=128, unet=None)

Module which detects and described local features in an image using the DISK method. See [TFT20] for details.

Parameters:
• desc_dim (int, optional) – The dimension of the descriptor. Default: 128

• unet (, optional) – The U-Net to use. If None, a default U-Net is used. Kornia doesn’t provide the training code for DISK so this is only useful when using a custom checkpoint trained using the code released with the paper. The unet should take as input a tensor of shape $$(B, C, H, W)$$ and output a tensor of shape $$(B, \mathrm{desc\_dim} + 1, H, W)$$. Default: None

Example

>>> disk = DISK.from_pretrained('depth')
>>> images = torch.rand(1, 3, 256, 256)
>>> features = disk(images)

forward(images, n=None, window_size=5, score_threshold=0.0, pad_if_not_divisible=False)

Detects features in an image, returning keypoint locations, descriptors and detection scores.

Parameters:
• images (Tensor) – The image to detect features in. Shape $$(B, 3, H, W)$$.

• n (, optional) – The maximum number of keypoints to detect. If None, all keypoints are returned. Default: None

• window_size (int, optional) – The size of the non-maxima suppression window used to filter detections. Default: 5

• score_threshold (float, optional) – The minimum score a detection must have to be returned. See DISKFeatures for details. Default: 0.0

• pad_if_not_divisible (bool, optional) – if True, the non-16 divisible input is zero-padded to the closest 16-multiply Default: False

Return type:

Returns:

A list of length $$B$$ containing the detected features.

classmethod from_pretrained(checkpoint='depth', device=torch.device('cpu'))

Loads a pretrained model.

Depth model was trained using depth map supervision and is slightly more precise but biased to detect keypoints only where SfM depth is available. Epipolar model was trained using epipolar geometry supervision and is less precise but detects keypoints everywhere where they are matchable. The difference is especially pronounced on thin structures and on edges of objects.

Parameters:
• checkpoint (str, optional) – The checkpoint to load. One of ‘depth’ or ‘epipolar’. Default: "depth"

• device (device, optional) – The device to load the model to. Default: torch.device("cpu")

Return type:

DISK

Returns:

The pretrained model.

heatmap_and_dense_descriptors(images)

Returns the heatmap and the dense descriptors.

Parameters:

images (Tensor) – The image to detect features in. Shape $$(B, 3, H, W)$$.

Return type:

Returns:

A tuple of dense detection scores and descriptors. Shapes are $$(B, 1, H, W)$$ and $$(B, D, H, W)$$, where $$D$$ is the descriptor dimension.

class kornia.feature.DISKFeatures(keypoints, descriptors, detection_scores)

A data structure holding DISK keypoints, descriptors and detection scores for an image. Since DISK detects a varying number of keypoints per image, DISKFeatures is not batched.

Parameters:
• keypoints (Tensor) – Tensor of shape $$(N, 2)$$, where $$N$$ is the number of keypoints.

• descriptors (Tensor) – Tensor of shape $$(N, D)$$, where $$D$$ is the descriptor dimension.

• detection_scores (Tensor) – Tensor of shape $$(N,)$$ where the detection score can be interpreted as the log-probability of keeping a keypoint after it has been proposed (see the paper section Method → Feature distribution for details).

to(*args, **kwargs)

Calls torch.Tensor.to() on each tensor to move the keypoints, descriptors and detection scores to the specified device and/or data type.

Parameters:
• *args (Any) – Arguments passed to torch.Tensor.to().

• **kwargs (Any) – Keyword arguments passed to torch.Tensor.to().

Return type:

DISKFeatures

Returns:

A new DISKFeatures object with tensors of appropriate type and location.

property x: Tensor

Accesses the x coordinates of keypoints (along image width).

property y: Tensor

Accesses the y coordinates of keypoints (along image height).

class kornia.feature.SIFTFeature(num_features=8000, upright=False, rootsift=True, device=torch.device('cpu'), config=get_default_detector_config())

Convenience module, which implements DoG detector + (Root)SIFT descriptor.

Using kornia.feature.MultiResolutionDetector without blur pyramid Still not as good as OpenCV/VLFeat because of https://github.com/kornia/kornia/pull/884, but we are working on it

Parameters:
• img (Tensor) – image to extract features with shape $$(B,C,H,W)$$.

• mask (, optional) – a mask with weights where to apply the response function. The shape must be the same as the input image. Default: None

Return type:

Returns:

• Detected local affine frames with shape $$(B,N,2,3)$$.

• Response function values for corresponding lafs with shape $$(B,N,1)$$.

• Local descriptors of shape $$(B,N,D)$$ where $$D$$ is descriptor size.

class kornia.feature.SIFTFeatureScaleSpace(num_features=8000, upright=False, rootsift=True, device=torch.device('cpu'))

Convenience module, which implements DoG detector + (Root)SIFT descriptor. Using kornia.feature.ScaleSpaceDetector with blur pyramid.

Still not as good as OpenCV/VLFeat because of https://github.com/kornia/kornia/pull/884, but we are working on it

Parameters:
• img (Tensor) – image to extract features with shape $$(B,C,H,W)$$.

• mask (, optional) – a mask with weights where to apply the response function. The shape must be the same as the input image. Default: None

Return type:

Returns:

• Detected local affine frames with shape $$(B,N,2,3)$$.

• Response function values for corresponding lafs with shape $$(B,N,1)$$.

• Local descriptors of shape $$(B,N,D)$$ where $$D$$ is descriptor size.

class kornia.feature.GFTTAffNetHardNet(num_features=8000, upright=False, device=torch.device('cpu'), config=get_default_detector_config())

Convenience module, which implements GFTT detector + AffNet-HardNet descriptor.

Parameters:
• img (Tensor) – image to extract features with shape $$(B,C,H,W)$$.

• mask (, optional) – a mask with weights where to apply the response function. The shape must be the same as the input image. Default: None

Return type:

Returns:

• Detected local affine frames with shape $$(B,N,2,3)$$.

• Response function values for corresponding lafs with shape $$(B,N,1)$$.

• Local descriptors of shape $$(B,N,D)$$ where $$D$$ is descriptor size.

class kornia.feature.KeyNetAffNetHardNet(num_features=8000, upright=False, device=torch.device('cpu'), scale_laf=1.0)

Convenience module, which implements KeyNet detector + AffNet + HardNet descriptor.

Parameters:
• img (Tensor) – image to extract features with shape $$(B,C,H,W)$$.

• mask (, optional) – a mask with weights where to apply the response function. The shape must be the same as the input image. Default: None

Return type:

Returns:

• Detected local affine frames with shape $$(B,N,2,3)$$.

• Response function values for corresponding lafs with shape $$(B,N,1)$$.

• Local descriptors of shape $$(B,N,D)$$ where $$D$$ is descriptor size.

class kornia.feature.KeyNetHardNet(num_features=8000, upright=False, device=torch.device('cpu'), scale_laf=1.0)

Convenience module, which implements KeyNet detector + HardNet descriptor.

Parameters:
• img (Tensor) – image to extract features with shape $$(B,C,H,W)$$.

• mask (, optional) – a mask with weights where to apply the response function. The shape must be the same as the input image. Default: None

Return type:

Returns:

• Detected local affine frames with shape $$(B,N,2,3)$$.

• Response function values for corresponding lafs with shape $$(B,N,1)$$.

• Local descriptors of shape $$(B,N,D)$$ where $$D$$ is descriptor size.

## Matching¶

kornia.feature.match_nn(desc1, desc2, dm=None)

Function, which finds nearest neighbors in desc2 for each vector in desc1.

If the distance matrix dm is not provided, torch.cdist() is used.

Parameters:
• desc1 (Tensor) – Batch of descriptors of a shape $$(B1, D)$$.

• desc2 (Tensor) – Batch of descriptors of a shape $$(B2, D)$$.

• dm (, optional) – Tensor containing the distances from each descriptor in desc1 to each descriptor in desc2, shape of $$(B1, B2)$$. Default: None

Return type:

Returns:

• Descriptor distance of matching descriptors, shape of $$(B1, 1)$$.

• Long tensor indexes of matching descriptors in desc1 and desc2, shape of $$(B1, 2)$$.

kornia.feature.match_mnn(desc1, desc2, dm=None)

Function, which finds mutual nearest neighbors in desc2 for each vector in desc1.

If the distance matrix dm is not provided, torch.cdist() is used.

Parameters:
• desc1 (Tensor) – Batch of descriptors of a shape $$(B1, D)$$.

• desc2 (Tensor) – Batch of descriptors of a shape $$(B2, D)$$.

• dm (, optional) – Tensor containing the distances from each descriptor in desc1 to each descriptor in desc2, shape of $$(B1, B2)$$. Default: None

Return type:

Returns:

• Descriptor distance of matching descriptors, shape of. $$(B3, 1)$$.

• Long tensor indexes of matching descriptors in desc1 and desc2, shape of $$(B3, 2)$$, where 0 <= B3 <= min(B1, B2)

kornia.feature.match_snn(desc1, desc2, th=0.8, dm=None)

Function, which finds nearest neighbors in desc2 for each vector in desc1.

The method satisfies first to second nearest neighbor distance <= th.

If the distance matrix dm is not provided, torch.cdist() is used.

Parameters:
• desc1 (Tensor) – Batch of descriptors of a shape $$(B1, D)$$.

• desc2 (Tensor) – Batch of descriptors of a shape $$(B2, D)$$.

• th (float, optional) – distance ratio threshold. Default: 0.8

• dm (, optional) – Tensor containing the distances from each descriptor in desc1 to each descriptor in desc2, shape of $$(B1, B2)$$. Default: None

Return type:

Returns:

• Descriptor distance of matching descriptors, shape of $$(B3, 1)$$.

• Long tensor indexes of matching descriptors in desc1 and desc2. Shape: $$(B3, 2)$$, where 0 <= B3 <= B1.

kornia.feature.match_smnn(desc1, desc2, th=0.95, dm=None)

Function, which finds mutual nearest neighbors in desc2 for each vector in desc1.

the method satisfies first to second nearest neighbor distance <= th.

If the distance matrix dm is not provided, torch.cdist() is used.

Parameters:
• desc1 (Tensor) – Batch of descriptors of a shape $$(B1, D)$$.

• desc2 (Tensor) – Batch of descriptors of a shape $$(B2, D)$$.

• th (float, optional) – distance ratio threshold. Default: 0.95

• dm (, optional) – Tensor containing the distances from each descriptor in desc1 to each descriptor in desc2, shape of $$(B1, B2)$$. Default: None

Return type:

Returns:

• Descriptor distance of matching descriptors, shape of. $$(B3, 1)$$.

• Long tensor indexes of matching descriptors in desc1 and desc2, shape of $$(B3, 2)$$ where 0 <= B3 <= B1.

kornia.feature.match_fginn(desc1, desc2, lafs1, lafs2, th=0.8, spatial_th=10.0, mutual=False, dm=None)

Function, which finds nearest neighbors in desc2 for each vector in desc1.

The method satisfies first to second nearest neighbor distance <= th, and assures 2nd nearest neighbor is geometrically inconsistent with the 1st one (see [MMP15] for more details)

If the distance matrix dm is not provided, torch.cdist() is used.

Parameters:
• desc1 (Tensor) – Batch of descriptors of a shape $$(B1, D)$$.

• desc2 (Tensor) – Batch of descriptors of a shape $$(B2, D)$$.

• lafs1 (Tensor) – LAFs of a shape $$(1, B1, 2, 3)$$.

• lafs2 (Tensor) – LAFs of a shape $$(1, B2, 2, 3)$$.

• th (float, optional) – distance ratio threshold. Default: 0.8

• spatial_th (float, optional) – minimal distance in pixels to 2nd nearest neighbor. Default: 10.0

• mutual (bool, optional) – also perform mutual nearest neighbor check Default: False

• dm (, optional) – Tensor containing the distances from each descriptor in desc1 to each descriptor in desc2, shape of $$(B1, B2)$$. Default: None

Return type:

Returns:

• Descriptor distance of matching descriptors, shape of $$(B3, 1)$$.

• Long tensor indexes of matching descriptors in desc1 and desc2. Shape: $$(B3, 2)$$, where 0 <= B3 <= B1.

kornia.feature.match_adalam(desc1, desc2, lafs1, lafs2, config=None, hw1=None, hw2=None, dm=None)

Function, which performs descriptor matching, followed by AdaLAM filtering (see [CLO+20] for more details)

If the distance matrix dm is not provided, torch.cdist() is used.

Parameters:
• desc1 (Tensor) – Batch of descriptors of a shape $$(B1, D)$$.

• desc2 (Tensor) – Batch of descriptors of a shape $$(B2, D)$$.

• lafs1 (Tensor) – LAFs of a shape $$(1, B1, 2, 3)$$.

• lafs2 (Tensor) – LAFs of a shape $$(1, B2, 2, 3)$$.

• config (Optional[AdalamConfig], optional) – dict with AdaLAM config Default: None

• dm (, optional) – Tensor containing the distances from each descriptor in desc1 to each descriptor in desc2, shape of $$(B1, B2)$$. Default: None

Return type:

Returns:

• Descriptor distance of matching descriptors, shape of $$(B3, 1)$$.

• Long tensor indexes of matching descriptors in desc1 and desc2. Shape: $$(B3, 2)$$, where 0 <= B3 <= B1.

class kornia.feature.DescriptorMatcher(match_mode='snn', th=0.8)

Module version of matching functions.

See match_nn(), match_snn(),

match_mnn() or match_smnn() for more details.

Parameters:
• match_mode (str, optional) – type of matching, can be nn, snn, mnn, smnn. Default: "snn"

• th (float, optional) – threshold on distance ratio, or other quality measure. Default: 0.8

forward(desc1, desc2)
Parameters:
• desc1 (Tensor) – Batch of descriptors of a shape $$(B1, D)$$.

• desc2 (Tensor) – Batch of descriptors of a shape $$(B2, D)$$.

• lafs1 – LAFs of a shape $$(1, B1, 2, 3)$$.

• lafs2 – LAFs of a shape $$(1, B2, 2, 3)$$.

Return type:

Returns:

• Descriptor distance of matching descriptors, shape of $$(B3, 1)$$.

• Long tensor indexes of matching descriptors in desc1 and desc2,

shape of $$(B3, 2)$$ where $$0 <= B3 <= B1$$.

class kornia.feature.GeometryAwareDescriptorMatcher(match_mode='fginn', params={})

Module version of matching functions.

See match_nn(), match_snn(),

match_mnn() or match_smnn() for more details.

Parameters:
• match_mode (str, optional) – type of matching, can be fginn. Default: "fginn"

• th – threshold on distance ratio, or other quality measure.

forward(desc1, desc2, lafs1, lafs2)
Parameters:
Return type:

Returns:

• Descriptor distance of matching descriptors, shape of $$(B3, 1)$$.

• Long tensor indexes of matching descriptors in desc1 and desc2,

shape of $$(B3, 2)$$ where $$0 <= B3 <= B1$$.

class kornia.feature.LocalFeatureMatcher(local_feature, matcher)

Module, which finds correspondences between two images based on local features.

Parameters:
Returns:

Dictionary with image correspondences and confidence scores.

Return type:

Dict[str, Tensor]

Example

>>> img1 = torch.rand(1, 1, 320, 200)
>>> img2 = torch.rand(1, 1, 128, 128)
>>> input = {"image0": img1, "image1": img2}
>>> gftt_hardnet_matcher = LocalFeatureMatcher(
...     GFTTAffNetHardNet(10), kornia.feature.DescriptorMatcher('snn', 0.8)
... )
>>> out = gftt_hardnet_matcher(input)

forward(data)
Parameters:

data () – dictionary containing the input data in the following format:

Keyword Arguments:
• image0 – left image with shape $$(N, 1, H1, W1)$$.

• image1 – right image with shape $$(N, 1, H2, W2)$$.

• mask0 (optional) – left image mask. ‘0’ indicates a padded position $$(N, H1, W1)$$.

• mask1 (optional) – right image mask. ‘0’ indicates a padded position $$(N, H2, W2)$$.

Return type:

Returns:

• keypoints0, matching keypoints from image0 $$(NC, 2)$$.

• keypoints1, matching keypoints from image1 $$(NC, 2)$$.

• confidence, confidence score [0, 1] $$(NC)$$.

• lafs0, matching LAFs from image0 $$(1, NC, 2, 3)$$.

• lafs1, matching LAFs from image1 $$(1, NC, 2, 3)$$.

• batch_indexes, batch indexes for the keypoints and lafs $$(NC)$$.

class kornia.feature.LightGlueMatcher(feature_name='disk', params={})

LightGlue-based matcher in kornia API.

This is based on the original code from paper “LightGlue: Local Feature Matching at Light Speed”. See [LSP23] for more details.

Parameters:
• feature_name (str, optional) – type of feature for matching, can be disk or superpoint. Default: "disk"

• params (Dict, optional) – LightGlue params. Default: {}

forward(desc1, desc2, lafs1, lafs2, hw1=None, hw2=None)
Parameters:
Return type:

Returns:

• Descriptor distance of matching descriptors, shape of $$(B3, 1)$$.

• Long tensor indexes of matching descriptors in desc1 and desc2,

shape of $$(B3, 2)$$ where $$0 <= B3 <= B1$$.

class kornia.feature.LightGlue(features='superpoint', **conf_)
forward(data)

Match keypoints and descriptors between two images.

Return type:

dict

Input (dict):
image0: dict

keypoints: [B x M x 2] descriptors: [B x M x D] image: [B x C x H x W] or image_size: [B x 2]

image1: dict

keypoints: [B x N x 2] descriptors: [B x N x D] image: [B x C x H x W] or image_size: [B x 2]

Output (dict):

log_assignment: [B x M+1 x N+1] matches0: [B x M] matching_scores0: [B x M] matches1: [B x N] matching_scores1: [B x N] matches: List[[Si x 2]], scores: List[[Si]]

class kornia.feature.LoFTR(pretrained='outdoor', config=default_cfg)

Module, which finds correspondences between two images.

This is based on the original code from paper “LoFTR: Detector-Free Local Feature Matching with Transformers”. See [SSW+21] for more details.

If the distance matrix dm is not provided, torch.cdist() is used.

Parameters:
• config (, optional) – Dict with initialization parameters. Do not pass it, unless you know what you are doing. Default: default_cfg

• pretrained (, optional) – Download and set pretrained weights to the model. Options: ‘outdoor’, ‘indoor’. ‘outdoor’ is trained on the MegaDepth dataset and ‘indoor’ on the ScanNet. Default: "outdoor"

Returns:

Dictionary with image correspondences and confidence scores.

Example

>>> img1 = torch.rand(1, 1, 320, 200)
>>> img2 = torch.rand(1, 1, 128, 128)
>>> input = {"image0": img1, "image1": img2}
>>> loftr = LoFTR('outdoor')
>>> out = loftr(input)

forward(data)
Parameters:

data () – dictionary containing the input data in the following format:

Keyword Arguments:
• image0 – left image with shape $$(N, 1, H1, W1)$$.

• image1 – right image with shape $$(N, 1, H2, W2)$$.

• mask0 (optional) – left image mask. ‘0’ indicates a padded position $$(N, H1, W1)$$.

• mask1 (optional) – right image mask. ‘0’ indicates a padded position $$(N, H2, W2)$$.

Return type:

Returns:

• keypoints0, matching keypoints from image0 $$(NC, 2)$$.

• keypoints1, matching keypoints from image1 $$(NC, 2)$$.

• confidence, confidence score [0, 1] $$(NC)$$.

• batch_indexes, batch indexes for the keypoints and lafs $$(NC)$$.

class kornia.feature.OnnxLightGlue(weights=None, device='cpu')

Wrapper for loading LightGlue-ONNX models and running inference via ONNXRuntime.

LightGlue [LSP23] performs fast descriptor-based deep keypoint matching. This module requires onnxruntime to be installed.

If you have trained your own LightGlue model, see https://github.com/fabio-sim/LightGlue-ONNX for how to export the model to ONNX and optimize it.

Parameters:
• weights (, optional) – Pretrained weights, or a path to your own exported ONNX model. Available pretrained weights are 'disk', 'superpoint', 'disk_fp16', and 'superpoint_fp16'. Note that FP16 requires CUDA. Defaults to 'disk_fp16' if device is CUDA, and 'disk' if CPU. Default: None

• device (, optional) – Device to run inference on. Default: "cpu"

forward(data)

Match keypoints and descriptors between two images.

The output contains the matches (the indices of the matching keypoint pairs between the first and second image) and the corresponding confidence scores. Only a batch size of 1 is supported.

Parameters:

data () – Dictionary containing both images and the keypoints and descriptors thereof.

Return type:

Returns:

Dictionary containing the matches and scores.

data (dict):
image0 (dict):

keypoints (float32): $$(1, M, 2)$$

descriptors (float32): $$(1, M, D)$$

image: $$(1, C, H, W)$$ or image_size: $$(1, 2)$$

image1 (dict):

keypoints (float32): $$(1, N, 2)$$

descriptors (float32): $$(1, N, D)$$

image: $$(1, C, H, W)$$ or image_size: $$(1, 2)$$

output (dict):

matches (int64): $$(S, 2)$$

scores (float32): $$(S)$$

## Local Affine Frames (LAF)¶

kornia.feature.extract_patches_from_pyramid(img, laf, PS=32, normalize_lafs_before_extraction=True)

Extract patches defined by LAFs from image tensor.

Patches are extracted from appropriate pyramid level.

Parameters:
Return type:

Tensor

Returns:

patches with shape $$(B, N, CH, PS,PS)$$.

kornia.feature.extract_patches_simple(img, laf, PS=32, normalize_lafs_before_extraction=True)

Extract patches defined by LAFs from image tensor.

No smoothing applied, huge aliasing (better use extract_patches_from_pyramid).

Parameters:
Return type:

Tensor

Returns:

patches with shape $$(B, N, CH, PS,PS)$$.

kornia.feature.normalize_laf(LAF, images)

Normalize LAFs to [0,1] scale from pixel scale. See below:

B,N,H,W = images.size() MIN_SIZE = min(H - 1, W -1) [a11 a21 x] [a21 a22 y] becomes: [a11/MIN_SIZE a21/MIN_SIZE x/(W-1)] [a21/MIN_SIZE a22/MIN_SIZE y/(H-1)]

Parameters:
Returns:

$$(B, N, 2, 3)$$, scale in image percentage (0, 1)

Return type:

the denormalized LAF

kornia.feature.denormalize_laf(LAF, images)

De-normalize LAFs from scale to image scale. The convention is that center of 5-pixel image (coordinates from 0 to 4) is 2, and not 2.5.

B,N,H,W = images.size() MIN_SIZE = min(H - 1, W -1) [a11 a21 x] [a21 a22 y] becomes [a11*MIN_SIZE a21*MIN_SIZE x*(W-1)] [a21*MIN_SIZE a22*MIN_SIZE y*(W-1)]

Parameters:
Returns:

$$(B, N, 2, 3)$$, scale in pixels

Return type:

the denormalized LAF

kornia.feature.laf_to_boundary_points(LAF, n_pts=50)

Convert LAFs to boundary points of the regions + center.

Used for local features visualization, see visualize_laf function.

Parameters:
Returns:

$$(B, N, n_pts, 2)$$

Return type:

tensor of boundary points LAF

kornia.feature.ellipse_to_laf(ells)

Convert ellipse regions to LAF format.

Ellipse (a, b, c) and upright covariance matrix [a11 a12; 0 a22] are connected by inverse matrix square root: A = invsqrt([a b; b c]).

Parameters:

ells (Tensor) – tensor $$(B, N, 5)$$ of ellipses in Oxford format [x y a b c].

Return type:

Tensor

Returns:

LAF $$(B, N, 2, 3)$$

Example

>>> input = torch.ones(1, 10, 5)  # BxNx5
>>> output = ellipse_to_laf(input)  #  BxNx2x3

kornia.feature.make_upright(laf, eps=1e-9)

Rectify the affine matrix, so that it becomes upright.

Parameters:
Returns:

$$(B, N, 2, 3)$$

Return type:

laf

Example

>>> input = torch.ones(1, 5, 2, 3)  # BxNx2x3
>>> output = make_upright(input)  #  BxNx2x3

kornia.feature.scale_laf(laf, scale_coef)

Multiplies region part of LAF ([:, :, :2, :2]) by a scale_coefficient.

So the center, shape and orientation of the local feature stays the same, but the region area changes.

:param LAF $$: :type LAF :math:$$: B, N, 2, 3 :type scale_coef: :param scale_coef: broadcastable tensor or float.

Return type:

Tensor

Returns:

LAF $$(B, N, 2, 3)$$

Example

>>> input = torch.ones(1, 5, 2, 3)  # BxNx2x3
>>> scale = 0.5
>>> output = scale_laf(input, scale)  # BxNx2x3

kornia.feature.get_laf_scale(LAF)

Return a scale of the LAFs.

Parameters:

LAF (Tensor) – $$(B, N, 2, 3)$$

Return type:

Tensor

Returns:

scale $$(B, N, 1, 1)$$

Example

>>> input = torch.ones(1, 5, 2, 3)  # BxNx2x3
>>> output = get_laf_scale(input)  # BxNx1x1

kornia.feature.get_laf_center(LAF)

Return a center (keypoint) of the LAFs. The convention is that center of 5-pixel image (coordinates from 0 to 4) is 2, and not 2.5.

Parameters:

LAF (Tensor) – $$(B, N, 2, 3)$$

Return type:

Tensor

Returns:

xy $$(B, N, 2)$$

Example

>>> input = torch.ones(1, 5, 2, 3)  # BxNx2x3
>>> output = get_laf_center(input)  # BxNx2

kornia.feature.rotate_laf(LAF, angles_degrees)

Apply additional rotation to the LAFs. Compared to set_laf_orientation, the resulting rotation is original LAF orientation plus angles_degrees.

Parameters:
Return type:

Tensor

Returns:

LAF oriented with angles $$(B, N, 2, 3)$$

kornia.feature.get_laf_orientation(LAF)

Return orientation of the LAFs, in degrees.

Parameters:

LAF (Tensor) – $$(B, N, 2, 3)$$

Return type:

Tensor

Returns:

angle in degrees $$(B, N, 1)$$

Example

>>> input = torch.ones(1, 5, 2, 3)  # BxNx2x3
>>> output = get_laf_orientation(input)  # BxNx1

kornia.feature.set_laf_orientation(LAF, angles_degrees)

Change the orientation of the LAFs.

Parameters:
Return type:

Tensor

Returns:

LAF oriented with angles $$(B, N, 2, 3)$$

kornia.feature.laf_from_center_scale_ori(xy, scale=None, ori=None)

Creates a LAF from keypoint center, scale and orientation.

Useful to create kornia LAFs from OpenCV keypoints.

Parameters:
• xy (Tensor) – $$(B, N, 2)$$.

• scale (, optional) – $$(B, N, 1, 1)$$. If not provided, scale = 1.0 is assumed Default: None

• degrees (angle in) – $$(B, N, 1)$$. If not provided orientation = 0 is assumed

Return type:

Tensor

Returns:

LAF $$(B, N, 2, 3)$$

kornia.feature.laf_is_inside_image(laf, images, border=0)

Check if the LAF is touching or partly outside the image boundary.

Returns the mask of LAFs, which are fully inside the image, i.e. valid.

Parameters:
Return type:

Tensor

Returns:

mask with shape $$(B, N)$$.

kornia.feature.laf_to_three_points(laf)

Convert local affine frame(LAF) to alternative representation: coordinates of LAF center, LAF-x unit vector, LAF-y unit vector.

Parameters:

laf (Tensor) – $$(B, N, 2, 3)$$.

Return type:

Tensor

Returns:

threepts $$(B, N, 2, 3)$$.

kornia.feature.laf_from_three_points(threepts)

Convert three points to local affine frame.

Order is (0,0), (0, 1), (1, 0).

Parameters:

threepts (Tensor) – $$(B, N, 2, 3)$$.

Return type:

Tensor

Returns:

laf $$(B, N, 2, 3)$$.

kornia.feature.KORNIA_CHECK_LAF(laf, raises=True)

Check whether a Local Affine Frame (laf) has a valid shape.

Parameters:
• laf (Tensor) – local affine frame tensor to evaluate.

• raises (bool, optional) – bool indicating whether an exception should be raised upon failure. Default: True

Raises:

Exception – if the input laf does not have a shape $$(B,N,2,3)$$ and raises is True.

Return type:

bool

Example

>>> lafs = torch.rand(2, 10, 2, 3)
>>> KORNIA_CHECK_LAF(lafs)
True

kornia.feature.perspective_transform_lafs(trans_01, lafs_1)

Function that applies perspective transformations to a set of local affine frames (LAFs).

Parameters:
Return type:

Tensor

Returns:

tensor of N-dimensional points of shape $$(B, N, 2, 3)$$.

Examples

>>> rng = torch.manual_seed(0)
>>> lafs_1 = torch.rand(2, 4, 2, 3)  # BxNx2x3
>>> lafs_1
tensor([[[[0.4963, 0.7682, 0.0885],
[0.1320, 0.3074, 0.6341]],

[[0.4901, 0.8964, 0.4556],
[0.6323, 0.3489, 0.4017]],

[[0.0223, 0.1689, 0.2939],
[0.5185, 0.6977, 0.8000]],

[[0.1610, 0.2823, 0.6816],
[0.9152, 0.3971, 0.8742]]],

[[[0.4194, 0.5529, 0.9527],
[0.0362, 0.1852, 0.3734]],

[[0.3051, 0.9320, 0.1759],
[0.2698, 0.1507, 0.0317]],

[[0.2081, 0.9298, 0.7231],
[0.7423, 0.5263, 0.2437]],

[[0.5846, 0.0332, 0.1387],
[0.2422, 0.8155, 0.7932]]]])
>>> trans_01 = torch.eye(3).repeat(2, 1, 1)  # Bx3x3
>>> trans_01.shape
torch.Size([2, 3, 3])
>>> lafs_0 = perspective_transform_lafs(trans_01, lafs_1)  # BxNx2x3

class kornia.feature.PassLAF(*args, **kwargs)

Dummy module to use instead of local feature orientation or affine shape estimator.

forward(laf, img)
Parameters:
Return type:

Tensor

Returns:

LAF, unchanged $$(B, N, 2, 3)$$

class kornia.feature.PatchAffineShapeEstimator(patch_size=19, eps=1e-10)

Module, which estimates the second moment matrix of the patch gradients.

The method determines the affine shape of the local feature as in .

Parameters:
• patch_size (int, optional) – the input image patch size. Default: 19

• eps (float, optional) – for safe division. Default: 1e-10

forward(patch)
Parameters:

patch (Tensor) – $$(B, 1, H, W)$$

Returns:

ellipse_shape $$(B, 1, 3)$$

Return type:

torch.Tensor

class kornia.feature.LAFAffineShapeEstimator(patch_size=32, affine_shape_detector=None, preserve_orientation=True)

Module, which extracts patches using input images and local affine frames (LAFs).

Then runs PatchAffineShapeEstimator on patches to estimate LAFs shape.

Then original LAF shape is replaced with estimated one. The original LAF orientation is not preserved, so it is recommended to first run LAFAffineShapeEstimator and then LAFOrienter,

Parameters:
• patch_size (int, optional) – the input image patch size. Default: 32

• affine_shape_detector (, optional) – Patch affine shape estimator, PatchAffineShapeEstimator. Default: None

• preserve_orientation (bool, optional) – if True, the original orientation is preserved. Default: True

forward(laf, img)
Parameters:
Returns:

$$(B, N, 2, 3)$$

Return type:

LAF_out

class kornia.feature.LAFOrienter(patch_size=32, num_angular_bins=36, angle_detector=None)

Module, which extracts patches using input images and local affine frames (LAFs).

Then runs PatchDominantGradientOrientation or OriNet on patches and then rotates the LAFs by the estimated angles

Parameters:
forward(laf, img)
Parameters:
Returns:

$$(B, N, 2, 3)$$

Return type:

LAF_out

class kornia.feature.PatchDominantGradientOrientation(patch_size=32, num_angular_bins=36, eps=1e-8)

Module, which estimates the dominant gradient orientation of the given patches, in radians.

Zero angle points towards right.

Parameters:
• patch_size (int, optional) – size of the (square) input patch. Default: 32

• num_angular_bins (int, optional) – number of histogram bins. Default: 36

• eps (float, optional) – for safe division, and arctan. Default: 1e-8

forward(patch)
Parameters:

patch (Tensor) – $$(B, 1, H, W)$$

Returns:

$$(B)$$

Return type:

class kornia.feature.OriNet(pretrained=False, eps=1e-8)

Network, which estimates the canonical orientation of the given 32x32 patches, in radians.

Zero angle points towards right. This is based on the original code from paper “Repeatability Is Not Enough: Learning Discriminative Affine Regions via Discriminability””. See [MRM18] for more details.

Parameters:
• pretrained (bool, optional) – Download and set pretrained weights to the model. Default: False

• eps (float, optional) – to avoid division by zero in atan2. Default: 1e-8

Returns:

Shape:
• Input: (B, 1, 32, 32)

• Output: (B)

Examples

>>> input = torch.rand(16, 1, 32, 32)
>>> orinet = OriNet()
>>> angle = orinet(input) # 16

forward(patch)
Parameters:

patch (Tensor) – $$(B, 1, H, W)$$

Returns:

$$(B)$$

Return type:

class kornia.feature.LAFAffNetShapeEstimator(pretrained=False, preserve_orientation=True)

Module, which extracts patches using input images and local affine frames (LAFs).

Then runs AffNet on patches to estimate LAFs shape. This is based on the original code from paper “Repeatability Is Not Enough: Learning Discriminative Affine Regions via Discriminability””. See [MRM18] for more details.

Then original LAF shape is replaced with estimated one. The original LAF orientation is not preserved, so it is recommended to first run LAFAffineShapeEstimator and then LAFOrienter.

Parameters:

pretrained (bool, optional) – Download and set pretrained weights to the model. Default: False

forward(laf, img)
Parameters:
Returns:

$$(B, N, 2, 3)$$

Return type:

LAF_out

## Layers¶

class kornia.feature.FilterResponseNorm2d(num_features, eps=1e-6, is_bias=True, is_scale=True, is_eps_leanable=False)

Feature Response Normalization layer from ‘Filter Response Normalization Layer: Eliminating Batch Dependence in the Training of Deep Neural Networks’, see [SK20] for more details.

$y = \gamma \times \frac{x}{\sqrt{\mathrm{E}[x^2]} + |\epsilon|} + \beta$
Parameters:
Returns:

Normalized features

Return type:

torch.Tensor

Shape:
• Input: $$(B, \text{num_features}, H, W)$$

• Output: $$(B, \text{num_features}, H, W)$$

class kornia.feature.TLU(num_features)

TLU layer from ‘Filter Response Normalization Layer: Eliminating Batch Dependence in the Training of Deep Neural Networks, see [SK20] for more details. $${\tau}$$ is learnable per channel.

$y = \max(x, {\tau})$
Parameters:

num_features (int) – number of channels

Returns:

torch.Tensor

Shape:
• Input: $$(B, \text{num_features}, H, W)$$

• Output: $$(B, \text{num_features}, H, W)$$

## Other¶

class kornia.feature.DeFMO(pretrained=False)

Module that disentangle a fast-moving object from the background and performs deblurring.

This is based on the original code from paper “DeFMO: Deblurring and Shape Recovery

of Fast Moving Objects”. See [ROF+21] for more details.

Parameters:

pretrained (bool, optional) – Download and set pretrained weights to the model. Default: false.

Returns:

Temporal super-resolution without background.

Shape:
• Input: (B, 6, H, W)

• Output: (B, S, 4, H, W)

Examples

>>> import kornia
>>> input = torch.rand(2, 6, 240, 320)
>>> defmo = kornia.feature.DeFMO()
>>> tsr_nobgr = defmo(input) # 2x24x4x240x320

forward(input_data)

Define the computation performed at every call.

Should be overridden by all subclasses. :rtype: Tensor

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module` instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.