File size: 8,677 Bytes
717b269
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
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
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
import torch
import torch.nn as nn 
import torch.nn.functional as F 
import math 
from ...utils.geometry import rot6d_to_rotmat, aa_to_rotmat
from typing import Optional

def make_linear_layers(feat_dims, relu_final=True, use_bn=False):
    layers = []
    for i in range(len(feat_dims)-1):
        layers.append(nn.Linear(feat_dims[i], feat_dims[i+1]))

        # Do not use ReLU for final estimation
        if i < len(feat_dims)-2 or (i == len(feat_dims)-2 and relu_final):
            if use_bn:
                layers.append(nn.BatchNorm1d(feat_dims[i+1]))
            layers.append(nn.ReLU(inplace=True))

    return nn.Sequential(*layers)

def make_conv_layers(feat_dims, kernel=3, stride=1, padding=1, bnrelu_final=True):
    layers = []
    for i in range(len(feat_dims)-1):
        layers.append(
            nn.Conv2d(
                in_channels=feat_dims[i],
                out_channels=feat_dims[i+1],
                kernel_size=kernel,
                stride=stride,
                padding=padding
                ))
        # Do not use BN and ReLU for final estimation
        if i < len(feat_dims)-2 or (i == len(feat_dims)-2 and bnrelu_final):
            layers.append(nn.BatchNorm2d(feat_dims[i+1]))
            layers.append(nn.ReLU(inplace=True))

    return nn.Sequential(*layers)

def make_deconv_layers(feat_dims, bnrelu_final=True):
    layers = []
    for i in range(len(feat_dims)-1):
        layers.append(
            nn.ConvTranspose2d(
                in_channels=feat_dims[i],
                out_channels=feat_dims[i+1],
                kernel_size=4,
                stride=2,
                padding=1,
                output_padding=0,
                bias=False))

        # Do not use BN and ReLU for final estimation
        if i < len(feat_dims)-2 or (i == len(feat_dims)-2 and bnrelu_final):
            layers.append(nn.BatchNorm2d(feat_dims[i+1]))
            layers.append(nn.ReLU(inplace=True))

    return nn.Sequential(*layers)

def sample_joint_features(img_feat, joint_xy):
    height, width = img_feat.shape[2:]
    x = joint_xy[:, :, 0] / (width - 1) * 2 - 1
    y = joint_xy[:, :, 1] / (height - 1) * 2 - 1
    grid = torch.stack((x, y), 2)[:, :, None, :]
    img_feat = F.grid_sample(img_feat, grid, align_corners=True)[:, :, :, 0]  # batch_size, channel_dim, joint_num
    img_feat = img_feat.permute(0, 2, 1).contiguous()  # batch_size, joint_num, channel_dim
    return img_feat

def perspective_projection(points: torch.Tensor,
                           translation: torch.Tensor,
                           focal_length: torch.Tensor,
                           camera_center: Optional[torch.Tensor] = None,
                           rotation: Optional[torch.Tensor] = None) -> torch.Tensor:
    """
    Computes the perspective projection of a set of 3D points.
    Args:
        points (torch.Tensor): Tensor of shape (B, N, 3) containing the input 3D points.
        translation (torch.Tensor): Tensor of shape (B, 3) containing the 3D camera translation.
        focal_length (torch.Tensor): Tensor of shape (B, 2) containing the focal length in pixels.
        camera_center (torch.Tensor): Tensor of shape (B, 2) containing the camera center in pixels.
        rotation (torch.Tensor): Tensor of shape (B, 3, 3) containing the camera rotation.
    Returns:
        torch.Tensor: Tensor of shape (B, N, 2) containing the projection of the input points.
    """
    batch_size = points.shape[0]
    if rotation is None:
        rotation = torch.eye(3, device=points.device, dtype=points.dtype).unsqueeze(0).expand(batch_size, -1, -1)
    if camera_center is None:
        camera_center = torch.zeros(batch_size, 2, device=points.device, dtype=points.dtype)
    # Populate intrinsic camera matrix K.
    K = torch.zeros([batch_size, 3, 3], device=points.device, dtype=points.dtype)
    K[:,0,0] = focal_length[:,0]
    K[:,1,1] = focal_length[:,1]
    K[:,2,2] = 1.
    K[:,:-1, -1] = camera_center
    # Transform points
    points = torch.einsum('bij,bkj->bki', rotation, points)
    points = points + translation.unsqueeze(1)

    # Apply perspective distortion
    projected_points = points / points[:,:,-1].unsqueeze(-1)

    # Apply camera intrinsics
    projected_points = torch.einsum('bij,bkj->bki', K, projected_points)

    return projected_points[:, :, :-1]

class DeConvNet(nn.Module):
    def __init__(self, feat_dim=768, upscale=4):
        super(DeConvNet, self).__init__()
        self.first_conv = make_conv_layers([feat_dim, feat_dim//2], kernel=1, stride=1, padding=0, bnrelu_final=False)
        self.deconv = nn.ModuleList([])
        for i in range(int(math.log2(upscale))+1):
            if i==0:
                self.deconv.append(make_deconv_layers([feat_dim//2, feat_dim//4]))
            elif i==1:
                self.deconv.append(make_deconv_layers([feat_dim//2, feat_dim//4, feat_dim//8]))
            elif i==2:
                self.deconv.append(make_deconv_layers([feat_dim//2, feat_dim//4, feat_dim//8, feat_dim//8]))

    def forward(self, img_feat):
        
        face_img_feats = []
        img_feat = self.first_conv(img_feat)
        face_img_feats.append(img_feat)
        for i, deconv in enumerate(self.deconv):
            scale = 2**i
            img_feat_i = deconv(img_feat)
            face_img_feat = img_feat_i
            face_img_feats.append(face_img_feat)
        return face_img_feats[::-1]   # high resolution -> low resolution

class DeConvNet_v2(nn.Module):
    def __init__(self, feat_dim=768):
        super(DeConvNet_v2, self).__init__()
        self.first_conv = make_conv_layers([feat_dim, feat_dim//2], kernel=1, stride=1, padding=0, bnrelu_final=False)
        self.deconv = nn.Sequential(*[nn.ConvTranspose2d(in_channels=feat_dim//2, out_channels=feat_dim//4, kernel_size=4, stride=4, padding=0, output_padding=0, bias=False), 
                       nn.BatchNorm2d(feat_dim//4), 
                       nn.ReLU(inplace=True)])
    
    def forward(self, img_feat):
        
        face_img_feats = []
        img_feat = self.first_conv(img_feat)
        img_feat = self.deconv(img_feat) 
        
        return [img_feat]
        
class RefineNet(nn.Module):
    def __init__(self, cfg, feat_dim=1280, upscale=3):
        super(RefineNet, self).__init__()
        #self.deconv     = DeConvNet_v2(feat_dim=feat_dim) 
        #self.out_dim    = feat_dim//4
        
        self.deconv     = DeConvNet(feat_dim=feat_dim, upscale=upscale)
        self.out_dim    = feat_dim//8  + feat_dim//4 + feat_dim//2 
        self.dec_pose   = nn.Linear(self.out_dim, 96) 
        self.dec_cam    = nn.Linear(self.out_dim, 3)
        self.dec_shape  = nn.Linear(self.out_dim, 10)
        
        self.cfg        = cfg
        self.joint_rep_type = cfg.MODEL.MANO_HEAD.get('JOINT_REP', '6d')
        self.joint_rep_dim = {'6d': 6, 'aa': 3}[self.joint_rep_type]
        
    def forward(self, img_feat, verts_3d, pred_cam, pred_mano_feats, focal_length):
        B = img_feat.shape[0]
        
        img_feats = self.deconv(img_feat)
        
        img_feat_sizes = [img_feat.shape[2] for img_feat in img_feats] 
        
        temp_cams  = [torch.stack([pred_cam[:, 1], pred_cam[:, 2],  
                                  2*focal_length[:, 0]/(img_feat_size * pred_cam[:, 0] +1e-9)],dim=-1) for img_feat_size in img_feat_sizes] 

        verts_2d   = [perspective_projection(verts_3d,
                                translation=temp_cams[i],
                                focal_length=focal_length / img_feat_sizes[i]) for i in range(len(img_feat_sizes))]
        
        vert_feats = [sample_joint_features(img_feats[i], verts_2d[i]).max(1).values  for i in range(len(img_feat_sizes))] 

        vert_feats = torch.cat(vert_feats, dim=-1)

        delta_pose  = self.dec_pose(vert_feats)
        delta_betas = self.dec_shape(vert_feats)
        delta_cam   = self.dec_cam(vert_feats)

        
        pred_hand_pose = pred_mano_feats['hand_pose'] + delta_pose
        pred_betas     = pred_mano_feats['betas']     + delta_betas 
        pred_cam       = pred_mano_feats['cam']       + delta_cam

        joint_conversion_fn = {
                '6d': rot6d_to_rotmat,
                'aa': lambda x: aa_to_rotmat(x.view(-1, 3).contiguous())
            }[self.joint_rep_type]
 
        pred_hand_pose = joint_conversion_fn(pred_hand_pose).view(B, self.cfg.MANO.NUM_HAND_JOINTS+1, 3, 3)
        
        pred_mano_params = {'global_orient': pred_hand_pose[:, [0]],
                            'hand_pose': pred_hand_pose[:, 1:],
                            'betas': pred_betas}
        
        return  pred_mano_params, pred_cam