kornia.feature

Detectors

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

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[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

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)[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

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]]]])
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

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)[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.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) – 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, num_spatial_bins, num_spatial_bins)

  • Output: (B, 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.

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.

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.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.

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

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

Tuple[Tensor, Tensor]

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

Tuple[Tensor, Tensor]

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

Tuple[Tensor, Tensor]

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

Tuple[Tensor, Tensor]

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.LoFTR(pretrained='outdoor', config={'backbone_type': 'ResNetFPN', 'coarse': {'attention': 'linear', 'd_ffn': 256, 'd_model': 256, 'layer_names': ['self', 'cross', 'self', 'cross', 'self', 'cross', 'self', 'cross'], 'nhead': 8, 'temp_bug_fix': False}, 'fine': {'attention': 'linear', 'd_ffn': 128, 'd_model': 128, 'layer_names': ['self', 'cross'], 'nhead': 8}, 'fine_concat_coarse_feat': True, 'fine_window_size': 5, 'match_coarse': {'border_rm': 2, 'dsmax_temperature': 0.1, 'match_type': 'dual_softmax', 'skh_init_bin_score': 1.0, 'skh_iters': 3, 'skh_prefilter': True, 'thr': 0.2, 'train_coarse_percent': 0.4, 'train_pad_num_gt_min': 200}, 'resnetfpn': {'block_dims': [128, 196, 256], 'initial_dim': 128}, 'resolution': (8, 2)})[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: {'backbone_type': 'ResNetFPN', 'coarse': {'attention': 'linear', 'd_ffn': 256, 'd_model': 256, 'layer_names': ['self', 'cross', 'self', 'cross', 'self', 'cross', 'self', 'cross'], 'nhead': 8, 'temp_bug_fix': False}, 'fine': {'attention': 'linear', 'd_ffn': 128, 'd_model': 128, 'layer_names': ['self', 'cross'], 'nhead': 8}, 'fine_concat_coarse_feat': True, 'fine_window_size': 5, 'match_coarse': {'border_rm': 2, 'dsmax_temperature': 0.1, 'match_type': 'dual_softmax', 'skh_init_bin_score': 1.0, 'skh_iters': 3, 'skh_prefilter': True, 'thr': 0.2, 'train_coarse_percent': 0.4, 'train_pad_num_gt_min': 200}, 'resnetfpn': {'block_dims': [128, 196, 256], 'initial_dim': 128}, 'resolution': (8, 2)}

  • 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)

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

Tensor

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

Tensor

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

Tensor

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

Tensor

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]).

See also https://github.com/vlfeat/vlfeat/blob/master/toolbox/sift/vl_frame2oell.m

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

Tensor

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

Tensor

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, ori)[source]

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

Parameters

  • xy (Tensor) – tensor [BxNx2].

  • scale (Tensor) – tensor [BxNx1x1].

  • ori (Tensor) – tensor [BxNx1].

Return type

Tensor

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

Tensor

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

Non Maxima Suppression

kornia.feature.non_maxima_suppression2d(input, kernel_size, mask_only=False)

Apply non maxima suppression to filter.

See NonMaximaSuppression2d for details.

Return type

Tensor

kornia.feature.non_maxima_suppression3d(input, kernel_size, mask_only=False)

Apply non maxima suppression to filter.

See NonMaximaSuppression3d for details.

Return type

Tensor

kornia.feature.nms2d(input, kernel_size, mask_only=False)[source]

Apply non maxima suppression to filter.

See NonMaximaSuppression2d for details.

Return type

Tensor

kornia.feature.nms3d(input, kernel_size, mask_only=False)[source]

Apply non maxima suppression to filter.

See NonMaximaSuppression3d for details.

Return type

Tensor

Module

class kornia.feature.NonMaximaSuppression2d(kernel_size)[source]

Apply non maxima suppression to filter.

class kornia.feature.NonMaximaSuppression3d(kernel_size)[source]

Apply non maxima suppression to filter.

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.ScaleSpaceDetector(num_features=500, mr_size=6.0, scale_pyr_module=ScalePyramid(n_levels=3, init_sigma=1.6, min_size=15, extra_levels=3, border=6, sigma_step=1.2599210498948732, double_image=False), resp_module=BlobHessiangrads_mode=sobel), nms_module=ConvSoftArgmax3d(kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), temperature=tensor(1.), normalized_coordinates=False, eps=1e-08, strict_maxima_bonus=0.0, 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: BlobHessiangrads_mode=sobel)

  • nms_module (Module, optional) – outputs per-patch coordinates of the response maxima. See ConvSoftArgmax3d for details. Default: ConvSoftArgmax3d(kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1), temperature=tensor(1.), normalized_coordinates=False, eps=1e-08, strict_maxima_bonus=0.0, 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.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)[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

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

Tensor

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) – Default: 32

  • num_angular_bins (int, optional) – Default: 36

  • eps (float, optional) – for safe division, and arctan. Default: 1e-08

forward(patch)[source]
Parameters

patch (Tensor) – (torch.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)[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

Tensor

Returns

laf_out shape [BxNx2x3]

class kornia.feature.DeFMO(pretrained=False)[source]

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)[source]

Defines the computation performed at every call.

Should be overridden by all subclasses.

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.

Return type

Tensor