kornia.feature #

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)[source]#

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 (Union[Tensor, float], 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[Tensor], 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

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]]]])

kornia.feature.hessian_response(input, grads_mode='sobel', sigmas=None)[source]#

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[Tensor], 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

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)[source]#

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 \((B, C, D-1, H, W)\).

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)[source]#

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: torch.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)[source]#

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)[source]#

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)[source]#

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)[source]#

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)[source]#

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 dimentionality, 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)[source]#

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)[source]#

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)[source]#

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

Internally uses get_laf_descriptors().

Parameters

patch_descriptor_module (Optional[Module], optional) – patch descriptor module, e.g. SIFTDescriptor or HardNet. Default: HardNet.
patch_size (int, optional) – patch size in pixels, which descriptor expects. Default: 32
grayscale_descriptor (bool, optional) – True if patch_descriptor expects single-channel image. Default: True

forward(img, lafs)[source]#

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 features with shape \((B,C,H,W)\).
lafs (Tensor) – local affine frames \((B,N,2,3)\).

Return type

Returns

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

kornia.feature.get_laf_descriptors(img, lafs, patch_descriptor, patch_size=32, grayscale_descriptor=True)[source]#

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

Parameters

img (Tensor) – image features with shape \((B,C,H,W)\).
lafs (Tensor) – local affine frames \((B,N,2,3)\).
patch_descriptor (Module) – patch descriptor module, e.g. SIFTDescriptor or HardNet.
patch_size (int, optional) – patch size in pixels, which descriptor expects. Default: 32
grayscale_descriptor (bool, optional) – True if patch_descriptor expects single-channel image. Default: True

Return type

Returns

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

Matching#

kornia.feature.match_nn(desc1, desc2, dm=None)[source]#

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], 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)[source]#

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], 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)[source]#

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], 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=0.8, dm=None)[source]#

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: 0.8
dm (Optional[Tensor], 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.

class kornia.feature.DescriptorMatcher(match_mode='snn', th=0.8)[source]#

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)[source]#

Parameters

desc1 (Tensor) – Batch of descriptors of a shape \((B1, D)\).
desc2 (Tensor) – Batch of descriptors of a shape \((B2, D)\).

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.LocalFeature(detector, descriptor)[source]#

Module, which combines local feature detector and descriptor.

Parameters

detector (Module) – the detection module.
descriptor (LAFDescriptor) – the descriptor module.

forward(img, mask=None)[source]#

Parameters

img (Tensor) – image to extract features with shape \((B,C,H,W)\).
mask (Optional[Tensor], 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.SIFTFeature(num_features=8000, upright=False, rootsift=True, device=torch.device('cpu'))[source]#

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

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

forward(img, mask=None)#

Parameters

img (Tensor) – image to extract features with shape \((B,C,H,W)\).
mask (Optional[Tensor], 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'))[source]#

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

forward(img, mask=None)#

Parameters

img (Tensor) – image to extract features with shape \((B,C,H,W)\).
mask (Optional[Tensor], 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'))[source]#

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

forward(img, mask=None)#

Parameters

img (Tensor) – image to extract features with shape \((B,C,H,W)\).
mask (Optional[Tensor], 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'))[source]#

Convenience module, which implements KeyNet detector + HardNet descriptor.

forward(img, mask=None)#

Parameters

img (Tensor) – image to extract features with shape \((B,C,H,W)\).
mask (Optional[Tensor], 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.LocalFeatureMatcher(local_feature, matcher)[source]#

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

Parameters

local_feature (Module) – Local feature detector. See GFTTAffNetHardNet.
matcher (Module) – Descriptor matcher, see DescriptorMatcher.

Returns

Dictionary with image correspondences and confidence scores.

Return type

Dict[str, torch.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)[source]#

Parameters

data (Dict[str, Tensor]) – 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

Dict[str, Tensor]

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.LoFTR(pretrained='outdoor', config=default_cfg)[source]#

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 (Dict, optional) – Dict with initiliazation parameters. Do not pass it, unless you know what you are doing`. Default: default_cfg
pretrained (Optional[str], 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)[source]#

Parameters

data (dict) – 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

Dict[str, Tensor]

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 \((NC)\).

Local Affine Frames (LAF)#

kornia.feature.extract_patches_from_pyramid(img, laf, PS=32, normalize_lafs_before_extraction=True)[source]#

Extract patches defined by LAFs from image tensor.

Patches are extracted from appropriate pyramid level.

Parameters

laf (Tensor) –
images – images, LAFs are detected in.
PS (int, optional) – patch size. Default: 32
normalize_lafs_before_extraction (bool, optional) – if True, lafs are normalized to image size. Default: True

Return type

Returns

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

kornia.feature.extract_patches_simple(img, laf, PS=32, normalize_lafs_before_extraction=True)[source]#

Extract patches defined by LAFs from image tensor.

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

Parameters

img (Tensor) – images, LAFs are detected in.
laf (Tensor) –
PS (int, optional) – patch size. Default: 32
normalize_lafs_before_extraction (bool, optional) – if True, lafs are normalized to image size. Default: True

Return type

Returns

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

kornia.feature.normalize_laf(LAF, images)[source]#

Normalize LAFs to [0,1] scale from pixel scale. See below:: B,N,H,W = images.size() MIN_SIZE = min(H,W) [a11 a21 x] [a21 a22 y] becomes: [a11/MIN_SIZE a21/MIN_SIZE x/W] [a21/MIN_SIZE a22/MIN_SIZE y/H]

Parameters

LAF (Tensor) – (torch.Tensor).
images (Tensor) – (torch.Tensor) images, LAFs are detected in

Returns

(torch.Tensor).

Return type

LAF

Shape:

Input: \((B, N, 2, 3)\)
Output: \((B, N, 2, 3)\)

kornia.feature.denormalize_laf(LAF, images)[source]#

De-normalize LAFs from scale to image scale.

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

Parameters

LAF (Tensor) –
images (Tensor) – images, LAFs are detected in.

Return type

Returns

the denormalized lafs.

Shape:

Input: \((B, N, 2, 3)\)
Output: \((B, N, 2, 3)\)

kornia.feature.laf_to_boundary_points(LAF, n_pts=50)[source]#

Convert LAFs to boundary points of the regions + center.

Used for local features visualization, see visualize_laf function.

Parameters

LAF (Tensor) –
n_pts (int, optional) – number of points to output. Default: 50

Return type

Returns

tensor of boundary points.

Shape:

Input: \((B, N, 2, 3)\)
Output: \((B, N, n_pts, 2)\)

kornia.feature.ellipse_to_laf(ells)[source]#

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 of ellipses in Oxford format [x y a b c].
Return type: Tensor
Returns: tensor of ellipses in LAF format.

Shape:

Input: \((B, N, 5)\)
Output: \((B, N, 2, 3)\)

Example

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

kornia.feature.make_upright(laf, eps=1e-09)[source]#

Rectify the affine matrix, so that it becomes upright.

Parameters

laf (Tensor) – tensor of LAFs.
eps (float, optional) – for safe division. Default: 1e-09

Return type

Returns

tensor of same shape.

Shape:

Input: \((B, N, 2, 3)\)
Output: \((B, N, 2, 3)\)

Example

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

kornia.feature.scale_laf(laf, scale_coef)[source]#

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.

Parameters

laf (Tensor) – tensor [BxNx2x3] or [BxNx2x2].
scale_coef (Union[float, Tensor]) – broadcastable tensor or float.

Return type

Returns

tensor BxNx2x3.

Shape:

Input: \((B, N, 2, 3)\)
Input: \((B, N,)\) or ()
Output: \((B, N, 1, 1)\)

Example

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

kornia.feature.get_laf_scale(LAF)[source]#

Return a scale of the LAFs.

Parameters: LAF (Tensor) – tensor [BxNx2x3] or [BxNx2x2].
Return type: Tensor
Returns: tensor BxNx1x1.

Shape:

Input: :math: (B, N, 2, 3)
Output: :math: (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)[source]#

Return a center (keypoint) of the LAFs.

Parameters: LAF (Tensor) – tensor [BxNx2x3].
Return type: Tensor
Returns: tensor BxNx2.

Shape:

Input: :math: (B, N, 2, 3)
Output: :math: (B, N, 2)

Example

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

kornia.feature.get_laf_orientation(LAF)[source]#

Return orientation of the LAFs, in degrees.

Parameters: LAF (Tensor) – (torch.Tensor): tensor [BxNx2x3].
Returns: tensor BxNx1 .
Return type: torch.Tensor

Shape:

Input: :math: (B, N, 2, 3)
Output: :math: (B, N, 1)

Example

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

kornia.feature.laf_from_center_scale_ori(xy, scale=None, ori=None)[source]#

Return orientation of the LAFs, in radians. Useful to create kornia LAFs from OpenCV keypoints.

Parameters

xy (Tensor) – tensor [BxNx2].
scale (Optional[Tensor], optional) – tensor [BxNx1x1]. If not provided, scale = 1 is assumed Default: None
ori (Optional[Tensor], optional) – tensor [BxNx1]. If not provided orientation = 0 is assumed Default: None

Return type

Returns

tensor BxNx2x3.

kornia.feature.laf_is_inside_image(laf, images, border=0)[source]#

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

laf (Tensor) – \((B, N, 2, 3)\).
images (Tensor) – images, lafs are detected in \((B, CH, H, W)\).
border (int, optional) – additional border. Default: 0

Return type

Returns

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

kornia.feature.laf_to_three_points(laf)[source]#

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

Parameters: laf (Tensor) – \((B, N, 2, 3)\).
Returns: threepts \((B, N, 2, 3)\).

kornia.feature.laf_from_three_points(threepts)[source]#

Convert three points to local affine frame.

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

Parameters: threepts (Tensor) – \((B, N, 2, 3)\).
Returns: laf \((B, N, 2, 3)\).

kornia.feature.raise_error_if_laf_is_not_valid(laf)[source]#

Auxiliary function, which verifies that input.

Parameters: laf (Tensor) – [BxNx2x3] shape.
Return type: None

kornia.feature.perspective_transform_lafs(trans_01, lafs_1)[source]#

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

Parameters

trans_01 (Tensor) – tensor for perspective transformations of shape \((B, 3, 3)\).
lafs_1 (Tensor) – tensor of lafs of shape \((B, N, 2, 3)\).

Return type

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

Module#

class kornia.feature.BlobHessian(grads_mode='sobel')[source]#

Module that calculates Hessian blobs.

See hessian_response() for details.

class kornia.feature.CornerGFTT(grads_mode='sobel')[source]#

Module that calculates Shi-Tomasi corners.

See gfft_response() for details.

class kornia.feature.CornerHarris(k, grads_mode='sobel')[source]#

Module that calculates Harris corners.

See harris_response() for details.

class kornia.feature.BlobDoG[source]#

Module that calculates Difference-of-Gaussians blobs.

See dog_response() for details.

class kornia.feature.KeyNet(pretrained=False, keynet_conf=keynet_config['KeyNet_default_config'])[source]#

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 (Dict, optional) – Dict with initiliazation parameters. Do not pass it, unless you know what you are doing`. Default: keynet_config['KeyNet_default_config']

Returns

KeyNet response score.

Shape:

Input: \((B, 1, H, W)\)
Output: \((B, 1, H, W)\)

class kornia.feature.FilterResponseNorm2d(num_features, eps=1e-06, is_bias=True, is_scale=True, is_eps_leanable=False)[source]#

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

num_features (int) – number of channels
eps (float, optional) – normalization constant Default: 1e-06
is_bias (bool, optional) – use bias Default: True
is_scale (bool, optional) – use scale Default: True
drop_rate – dropout rate,
is_eps_leanable (bool, optional) – if eps is learnable Default: False

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)[source]#

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

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)[source]#

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

forward(img, mask=None)[source]#

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[Tensor], 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_config['KeyNet_default_config'], ori_module=PassLAF(), aff_module=PassLAF())[source]#

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 (Dict, optional) – Dict with initiliazation parameters. Do not pass it, unless you know what you are doing`. Default: keynet_config['KeyNet_default_config']
ori_module (Module, optional) – for local feature orientation estimation. Default: PassLAF, which does nothing. See LAFOrienter for details.
aff_module (Module, optional) – for local feature affine shape estimation. Default: PassLAF, which does nothing. See LAFAffineShapeEstimator for details.

forward(img, mask=None)[source]#

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

class kornia.feature.PassLAF[source]#

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

forward(laf, img)[source]#

Parameters

laf (Tensor) – 4d tensor.
img (Tensor) – the input image tensor.

Returns

unchanged laf from the input.

Return type

torch.Tensor

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

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

The method determines the affine shape of the local feature as in [Baumberg00].

Parameters

patch_size (int, optional) – the input image patch size. Default: 19
eps (float, optional) – for safe division. Default: 1e-10

forward(patch)[source]#

Parameters: patch (Tensor) – (torch.Tensor) shape [Bx1xHxW]
Returns: ellipse_shape shape [Bx1x3]
Return type: torch.Tensor

class kornia.feature.LAFAffineShapeEstimator(patch_size=32, affine_shape_detector=None, preserve_orientation=True)[source]#

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[Module], optional) – Patch affine shape estimator, PatchAffineShapeEstimator. Default: None
preserve_orientation (bool, optional) – if True, the original orientation is preserved. Default: True

forward(laf, img)[source]#

Parameters

laf (Tensor) – (torch.Tensor) shape [BxNx2x3]
img (Tensor) – (torch.Tensor) shape [Bx1xHxW]

Returns

laf_out shape [BxNx2x3]

Return type

torch.Tensor

class kornia.feature.LAFOrienter(patch_size=32, num_angular_bins=36, angle_detector=None)[source]#

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

patch_size (int, optional) – Default: 32
num_angular_bins (int, optional) – Default: 36
angle_detector (Optional[Module], optional) – Patch orientation estimator, e.g. PatchDominantGradientOrientation or OriNet. Default: None

forward(laf, img)[source]#

Parameters

laf (Tensor) – shape [BxNx2x3]
img (Tensor) – shape [Bx1xHxW]

Return type

Returns

laf_out, shape [BxNx2x3]

class kornia.feature.PatchDominantGradientOrientation(patch_size=32, num_angular_bins=36, eps=1e-08)[source]#

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-08

forward(patch)[source]#

Parameters: patch (Tensor) – shape [Bx1xHxW]
Returns: angle shape [B]
Return type: torch.Tensor

class kornia.feature.OriNet(pretrained=False, eps=1e-08)[source]#

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-08

Returns

Angle in radians.

Shape:

Input: (B, 1, 32, 32)
Output: (B)

Examples

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

forward(patch)[source]#

Parameters: patch (Tensor) – (torch.Tensor) shape [Bx1xHxW]
Returns: (torch.Tensor) shape [B]
Return type: patch

class kornia.feature.LAFAffNetShapeEstimator(pretrained=False, preserve_orientation=True)[source]#

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)[source]#

Parameters

laf (Tensor) – shape [BxNx2x3]
img (Tensor) – shape [Bx1xHxW]

Return type