tps.py 10.7 KB
Newer Older
W
WenmuZhou 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104
# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import paddle
from paddle import nn, ParamAttr
from paddle.nn import functional as F
import numpy as np


class ConvBNLayer(nn.Layer):
    def __init__(self,
                 in_channels,
                 out_channels,
                 kernel_size,
                 stride=1,
                 groups=1,
                 act=None,
                 name=None):
        super(ConvBNLayer, self).__init__()
        self.conv = nn.Conv2D(
            in_channels=in_channels,
            out_channels=out_channels,
            kernel_size=kernel_size,
            stride=stride,
            padding=(kernel_size - 1) // 2,
            groups=groups,
            weight_attr=ParamAttr(name=name + "_weights"),
            bias_attr=False)
        bn_name = "bn_" + name
        self.bn = nn.BatchNorm(
            out_channels,
            act=act,
            param_attr=ParamAttr(name=bn_name + '_scale'),
            bias_attr=ParamAttr(bn_name + '_offset'),
            moving_mean_name=bn_name + '_mean',
            moving_variance_name=bn_name + '_variance')

    def forward(self, x):
        x = self.conv(x)
        x = self.bn(x)
        return x


class LocalizationNetwork(nn.Layer):
    def __init__(self, in_channels, num_fiducial, loc_lr, model_name):
        super(LocalizationNetwork, self).__init__()
        self.F = num_fiducial
        F = num_fiducial
        if model_name == "large":
            num_filters_list = [64, 128, 256, 512]
            fc_dim = 256
        else:
            num_filters_list = [16, 32, 64, 128]
            fc_dim = 64

        self.block_list = []
        for fno in range(0, len(num_filters_list)):
            num_filters = num_filters_list[fno]
            name = "loc_conv%d" % fno
            conv = self.add_sublayer(
                name,
                ConvBNLayer(
                    in_channels=in_channels,
                    out_channels=num_filters,
                    kernel_size=3,
                    act='relu',
                    name=name))
            self.block_list.append(conv)
            if fno == len(num_filters_list) - 1:
                pool = nn.AdaptiveAvgPool2D(1)
            else:
                pool = nn.MaxPool2D(kernel_size=2, stride=2, padding=0)
            in_channels = num_filters
            self.block_list.append(pool)
        name = "loc_fc1"
        self.fc1 = nn.Linear(
            in_channels,
            fc_dim,
            weight_attr=ParamAttr(
                learning_rate=loc_lr, name=name + "_w"),
            bias_attr=ParamAttr(name=name + '.b_0'),
            name=name)

        # Init fc2 in LocalizationNetwork
        initial_bias = self.get_initial_fiducials()
        initial_bias = initial_bias.reshape(-1)
        name = "loc_fc2"
        param_attr = ParamAttr(
            learning_rate=loc_lr,
W
WenmuZhou 已提交
105
            initializer=nn.initializer.Assign(np.zeros([fc_dim, F * 2])),
W
WenmuZhou 已提交
106 107 108
            name=name + "_w")
        bias_attr = ParamAttr(
            learning_rate=loc_lr,
W
WenmuZhou 已提交
109
            initializer=nn.initializer.Assign(initial_bias),
W
WenmuZhou 已提交
110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
            name=name + "_b")
        self.fc2 = nn.Linear(
            fc_dim,
            F * 2,
            weight_attr=param_attr,
            bias_attr=bias_attr,
            name=name)
        self.out_channels = F * 2

    def forward(self, x):
        """
           Estimating parameters of geometric transformation
           Args:
               image: input
           Return:
               batch_C_prime: the matrix of the geometric transformation
        """
        B = x.shape[0]
        i = 0
        for block in self.block_list:
            x = block(x)
W
WenmuZhou 已提交
131
        x = x.squeeze(axis=2).squeeze(axis=2)
W
WenmuZhou 已提交
132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178
        x = self.fc1(x)

        x = F.relu(x)
        x = self.fc2(x)
        x = x.reshape(shape=[-1, self.F, 2])
        return x

    def get_initial_fiducials(self):
        """ see RARE paper Fig. 6 (a) """
        F = self.F
        ctrl_pts_x = np.linspace(-1.0, 1.0, int(F / 2))
        ctrl_pts_y_top = np.linspace(0.0, -1.0, num=int(F / 2))
        ctrl_pts_y_bottom = np.linspace(1.0, 0.0, num=int(F / 2))
        ctrl_pts_top = np.stack([ctrl_pts_x, ctrl_pts_y_top], axis=1)
        ctrl_pts_bottom = np.stack([ctrl_pts_x, ctrl_pts_y_bottom], axis=1)
        initial_bias = np.concatenate([ctrl_pts_top, ctrl_pts_bottom], axis=0)
        return initial_bias


class GridGenerator(nn.Layer):
    def __init__(self, in_channels, num_fiducial):
        super(GridGenerator, self).__init__()
        self.eps = 1e-6
        self.F = num_fiducial

        name = "ex_fc"
        initializer = nn.initializer.Constant(value=0.0)
        param_attr = ParamAttr(
            learning_rate=0.0, initializer=initializer, name=name + "_w")
        bias_attr = ParamAttr(
            learning_rate=0.0, initializer=initializer, name=name + "_b")
        self.fc = nn.Linear(
            in_channels,
            6,
            weight_attr=param_attr,
            bias_attr=bias_attr,
            name=name)

    def forward(self, batch_C_prime, I_r_size):
        """
        Generate the grid for the grid_sampler.
        Args:
            batch_C_prime: the matrix of the geometric transformation
            I_r_size: the shape of the input image
        Return:
            batch_P_prime: the grid for the grid_sampler
        """
W
WenmuZhou 已提交
179 180 181 182 183 184 185
        C = self.build_C_paddle()
        P = self.build_P_paddle(I_r_size)

        inv_delta_C_tensor = self.build_inv_delta_C_paddle(C).astype('float32')
        # inv_delta_C_tensor = paddle.zeros((23,23)).astype('float32')
        P_hat_tensor = self.build_P_hat_paddle(
            C, paddle.to_tensor(P)).astype('float32')
W
WenmuZhou 已提交
186 187 188 189 190 191 192 193 194 195 196 197 198 199

        inv_delta_C_tensor.stop_gradient = True
        P_hat_tensor.stop_gradient = True

        batch_C_ex_part_tensor = self.get_expand_tensor(batch_C_prime)

        batch_C_ex_part_tensor.stop_gradient = True

        batch_C_prime_with_zeros = paddle.concat(
            [batch_C_prime, batch_C_ex_part_tensor], axis=1)
        batch_T = paddle.matmul(inv_delta_C_tensor, batch_C_prime_with_zeros)
        batch_P_prime = paddle.matmul(P_hat_tensor, batch_T)
        return batch_P_prime

W
WenmuZhou 已提交
200
    def build_C_paddle(self):
W
WenmuZhou 已提交
201 202
        """ Return coordinates of fiducial points in I_r; C """
        F = self.F
W
WenmuZhou 已提交
203 204 205 206 207 208
        ctrl_pts_x = paddle.linspace(-1.0, 1.0, int(F / 2))
        ctrl_pts_y_top = -1 * paddle.ones([int(F / 2)])
        ctrl_pts_y_bottom = paddle.ones([int(F / 2)])
        ctrl_pts_top = paddle.stack([ctrl_pts_x, ctrl_pts_y_top], axis=1)
        ctrl_pts_bottom = paddle.stack([ctrl_pts_x, ctrl_pts_y_bottom], axis=1)
        C = paddle.concat([ctrl_pts_top, ctrl_pts_bottom], axis=0)
W
WenmuZhou 已提交
209 210
        return C  # F x 2

W
WenmuZhou 已提交
211 212 213 214 215 216 217 218
    def build_P_paddle(self, I_r_size):
        I_r_height, I_r_width = I_r_size
        I_r_grid_x = (
            paddle.arange(-I_r_width, I_r_width, 2).astype('float32') + 1.0
        ) / I_r_width  # self.I_r_width
        I_r_grid_y = (
            paddle.arange(-I_r_height, I_r_height, 2).astype('float32') + 1.0
        ) / I_r_height  # self.I_r_height
W
WenmuZhou 已提交
219
        # P: self.I_r_width x self.I_r_height x 2
W
WenmuZhou 已提交
220 221
        P = paddle.stack(paddle.meshgrid(I_r_grid_x, I_r_grid_y), axis=2)
        P = paddle.transpose(P, perm=[1, 0, 2])
W
WenmuZhou 已提交
222 223 224
        # n (= self.I_r_width x self.I_r_height) x 2
        return P.reshape([-1, 2])

W
WenmuZhou 已提交
225
    def build_inv_delta_C_paddle(self, C):
W
WenmuZhou 已提交
226 227
        """ Return inv_delta_C which is needed to calculate T """
        F = self.F
W
WenmuZhou 已提交
228
        hat_C = paddle.zeros((F, F), dtype='float32')  # F x F
W
WenmuZhou 已提交
229 230
        for i in range(0, F):
            for j in range(i, F):
W
WenmuZhou 已提交
231 232 233 234 235 236 237 238
                if i == j:
                    hat_C[i, j] = 1
                else:
                    r = paddle.norm(C[i] - C[j])
                    hat_C[i, j] = r
                    hat_C[j, i] = r
        hat_C = (hat_C**2) * paddle.log(hat_C)
        delta_C = paddle.concat(  # F+3 x F+3
W
WenmuZhou 已提交
239
            [
W
WenmuZhou 已提交
240 241 242 243 244 245 246 247 248
                paddle.concat(
                    [paddle.ones((F, 1)), C, hat_C], axis=1),  # F x F+3
                paddle.concat(
                    [paddle.zeros((2, 3)), paddle.transpose(
                        C, perm=[1, 0])],
                    axis=1),  # 2 x F+3
                paddle.concat(
                    [paddle.zeros((1, 3)), paddle.ones((1, F))],
                    axis=1)  # 1 x F+3
W
WenmuZhou 已提交
249 250
            ],
            axis=0)
W
WenmuZhou 已提交
251
        inv_delta_C = paddle.inverse(delta_C)
W
WenmuZhou 已提交
252 253
        return inv_delta_C  # F+3 x F+3

W
WenmuZhou 已提交
254
    def build_P_hat_paddle(self, C, P):
W
WenmuZhou 已提交
255 256 257 258
        F = self.F
        eps = self.eps
        n = P.shape[0]  # n (= self.I_r_width x self.I_r_height)
        # P_tile: n x 2 -> n x 1 x 2 -> n x F x 2
W
WenmuZhou 已提交
259 260
        P_tile = paddle.tile(paddle.unsqueeze(P, axis=1), (1, F, 1))
        C_tile = paddle.unsqueeze(C, axis=0)  # 1 x F x 2
W
WenmuZhou 已提交
261 262
        P_diff = P_tile - C_tile  # n x F x 2
        # rbf_norm: n x F
W
WenmuZhou 已提交
263 264
        rbf_norm = paddle.norm(P_diff, p=2, axis=2, keepdim=False)

W
WenmuZhou 已提交
265
        # rbf: n x F
W
WenmuZhou 已提交
266 267 268
        rbf = paddle.multiply(
            paddle.square(rbf_norm), paddle.log(rbf_norm + eps))
        P_hat = paddle.concat([paddle.ones((n, 1)), P, rbf], axis=1)
W
WenmuZhou 已提交
269 270 271
        return P_hat  # n x F+3

    def get_expand_tensor(self, batch_C_prime):
W
WenmuZhou 已提交
272 273
        B, H, C = batch_C_prime.shape
        batch_C_prime = batch_C_prime.reshape([B, H * C])
W
WenmuZhou 已提交
274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
        batch_C_ex_part_tensor = self.fc(batch_C_prime)
        batch_C_ex_part_tensor = batch_C_ex_part_tensor.reshape([-1, 3, 2])
        return batch_C_ex_part_tensor


class TPS(nn.Layer):
    def __init__(self, in_channels, num_fiducial, loc_lr, model_name):
        super(TPS, self).__init__()
        self.loc_net = LocalizationNetwork(in_channels, num_fiducial, loc_lr,
                                           model_name)
        self.grid_generator = GridGenerator(self.loc_net.out_channels,
                                            num_fiducial)
        self.out_channels = in_channels

    def forward(self, image):
        image.stop_gradient = False
        batch_C_prime = self.loc_net(image)
W
WenmuZhou 已提交
291
        batch_P_prime = self.grid_generator(batch_C_prime, image.shape[2:])
W
WenmuZhou 已提交
292 293 294 295
        batch_P_prime = batch_P_prime.reshape(
            [-1, image.shape[2], image.shape[3], 2])
        batch_I_r = F.grid_sample(x=image, grid=batch_P_prime)
        return batch_I_r