changes

face_swapper/src/dataset.py

@@ -6,6 +6,7 @@
 from torch.utils.data import Dataset
 from torchvision import io, transforms
 
+from .helper import warp_tensor
 from .types import Batch
 
 
@@ -35,6 +36,7 @@ def compose_transforms() -> transforms:
 			transforms.ColorJitter(brightness = 0.2, contrast = 0.2, saturation = 0.2, hue = 0.1),
 			transforms.RandomAffine(4, translate = (0.01, 0.01), scale = (0.98, 1.02), shear = (1, 1)),
 			transforms.ToTensor(),
+			lambda temp: warp_tensor(temp.unsqueeze(0), '__vgg_face_hq__to__arcface_128_v2__').squeeze(0),
 			transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
 		])
 

face_swapper/src/helper.py

@@ -1,10 +1,32 @@
+import torch
 from torch import Tensor, nn
 
-from .types import EmbedderModule, Embedding, Padding
+from .types import AlignmentMatrices, EmbedderModule, Embedding, Padding
+
+ALIGNMENT_MATRICES: AlignmentMatrices =\
+{
+	'__vgg_face_hq__to__arcface_128_v2__': torch.tensor(
+		[
+			[ 1.01305414, -0.00140513, -0.00585911 ],
+			[ 0.00140513, 1.01305414, 0.11169602 ]
+		], dtype = torch.float32),
+	'__arcface_128_v2__to__arcface_112_v2__': torch.tensor(
+		[
+			[ 8.75000016e-01, -1.07193451e-08, 3.80446920e-10 ],
+			[ 1.07193451e-08, 8.75000016e-01, -1.25000007e-01 ]
+		], dtype = torch.float32)
+}
+
+
+def warp_tensor(input_tensor : Tensor, alignment_matrix : str) -> Tensor:
+	matrix = ALIGNMENT_MATRICES.get(alignment_matrix).repeat(input_tensor.shape[0], 1, 1)
+	grid = nn.functional.affine_grid(matrix.to(input_tensor.device), list(input_tensor.shape))
+	output_tensor = nn.functional.grid_sample(input_tensor, grid, align_corners = False, padding_mode = 'reflection')
+	return output_tensor
 
 
 def calc_embedding(embedder : EmbedderModule, input_tensor : Tensor, padding : Padding) -> Embedding:
-	crop_tensor = input_tensor[:, :, 15: 241, 15: 241]
+	crop_tensor = warp_tensor(input_tensor, '__arcface_128_v2__to__arcface_112_v2__')
 	crop_tensor = nn.functional.interpolate(crop_tensor, size = (112, 112), mode = 'area')
 	crop_tensor[:, :, :padding[0], :] = 0
 	crop_tensor[:, :, 112 - padding[1]:, :] = 0

face_swapper/src/types.py

@@ -1,4 +1,4 @@
-from typing import Any, Tuple, TypeAlias
+from typing import Any, Dict, Tuple, TypeAlias
 
 from torch import Tensor
 from torch.nn import Module
@@ -17,3 +17,5 @@
 MotionExtractorModule : TypeAlias = Module
 
 OptimizerConfig : TypeAlias = Any
+
+AlignmentMatrices : TypeAlias = Dict[str, Tensor]