def preprocess_scene(x_2d, scene_rgb, scene_depth, scene_K, scene_Tcw,
scene_ori_rgb):
rand_R = torch.eye(3).unsqueeze(0)
N, L, C, H, W = scene_rgb.shape
# scene_rgb = scene_rgb.view(N, L, C, H, W)
scene_depth = scene_depth.view(N * L, 1, H, W)
scene_K = scene_K.view(N * L, 3, 3)
scene_Tcw = scene_Tcw.view(N * L, 3, 4)
# generate 3D world position of scene
d = scene_depth.view(N * L, H * W, 1) # dim (N*L, H*W, 1)
X_3d = batched_pi_inv(scene_K, x_2d, d) # dim (N*L, H*W, 3)
Rwc, twc = batched_inv_pose(
R=scene_Tcw[:, :3, :3],
t=scene_Tcw[:, :3, 3].squeeze(-1)) # dim (N*L, 3, 3), （N, 3)
X_world = batched_transpose(Rwc.cuda(), twc.cuda(),
X_3d) # dim (N*L, H*W, 3)
X_world = X_world.view(N, L * H * W, 3) # dim (N, L*H*W, 3)
scene_center = torch.mean(X_world, dim=1) # dim (N, 3)
X_world -= scene_center.view(N, 1, 3)
X_world = batched_transpose(
rand_R.cuda().expand(N, 3, 3),
torch.zeros(1, 3, 1).cuda().expand(N, 3, 1),
X_world) # dim (N, L*H*W, 3), data augmentation
X_world = X_world.view(N, L, H, W,
3).permute(0, 1, 4, 2,
3).contiguous() # dim (N, L, 3, H, W)
scene_input = torch.cat((scene_rgb, X_world), dim=2)
return scene_input.cuda(), scene_ori_rgb.cuda(), X_world.cuda(), \
torch.gt(scene_depth, 1e-5).cuda().view(N, L, H, W), scene_center.cuda(), rand_R.expand(N, 3, 3).cuda()
def wrapping(I_a, I_b, d_a, K, R, t):
"""
Wrap image by providing depth, rotation and translation
:param I_a:
:param I_b:
:param d_a:
:param K:
:param R:
:param t:
:return:
"""
import banet_track.ba_module as module
H, W, C = I_a.shape
# I_a = torch.from_numpy(I_a.transpose((2, 0, 1))).cuda().view((1, C, H, W))
I_b = torch.from_numpy(I_b.transpose((2, 0, 1))).cuda().view((1, C, H, W))
d_a = torch.from_numpy(d_a).cuda().view((1, H * W))
K = torch.from_numpy(K).cuda().view(1, 3, 3)
R = torch.from_numpy(R).cuda().view(1, 3, 3)
t = torch.from_numpy(t).cuda().view(1, 3)
x_a = module.x_2d_coords_torch(1, H, W).view(1, H * W,
2).cuda() # dim: (N, H*W, 2)
X_a_3d = module.batched_pi_inv(K, x_a, d_a.view((1, H * W, 1)))
X_b_3d = module.batched_transpose(R, t, X_a_3d)
x_b_2d, _ = module.batched_pi(K, X_b_3d)
x_b_2d_out = x_b_2d.cpu().numpy()
x_b_2d = module.batched_x_2d_normalize(H, W,
x_b_2d).view(1, H, W,
2) # (N, H, W, 2)
wrap_img_b = module.batched_interp2d(I_b, x_b_2d)
return wrap_img_b.cpu().numpy().transpose((0, 2, 3, 1)).reshape(
(H, W, C)), x_b_2d_out.reshape(H, W, 2)
def preprocess(x_2d, scene_rgb, scene_depth, scene_K, scene_Tcw, scene_ori_rgb, scene_center=None):
N, L, C, H, W = scene_rgb.shape
_, _, _, ori_H, ori_W = scene_ori_rgb.shape
scene_rgb = scene_rgb.view(N * L, C, H, W)
scene_depth = scene_depth.view(N * L, 1, H, W)
scene_K = scene_K.view(N * L, 3, 3)
scene_Tcw = scene_Tcw.view(N * L, 3, 4)
# generate 3D world position of scene
d = scene_depth.view(N * L, H * W, 1) # dim (N*L, H*W, 1)
X_3d = batched_pi_inv(scene_K, x_2d, d) # dim (N*L, H*W, 3)
Rwc, twc = batched_inv_pose(R=scene_Tcw[:, :3, :3],
t=scene_Tcw[:, :3, 3].squeeze(-1)) # dim (N*L, 3, 3), （N, 3)
X_world = batched_transpose(Rwc.cuda(), twc.cuda(), X_3d) # dim (N*L, H*W, 3)
X_world = X_world.view(N, L * H * W, 3) # dim (N, L*H*W, 3)
if scene_center is None:
scene_center = torch.mean(X_world, dim=1) # dim (N, 3)
X_world -= scene_center.view(N, 1, 3)
X_world = X_world.view(N * L, H, W, 3).permute(0, 3, 1, 2).contiguous() # dim (N * L, 3, H, W)
scene_input = torch.cat((scene_rgb, X_world), dim=1)
return scene_input.cuda(), \
scene_ori_rgb.view((N*L, 3, ori_H, ori_W)).cuda(), \
X_world.cuda(), \
torch.gt(scene_depth, 1e-5).cuda().view(N * L, H, W), \
scene_center.cuda()
def compute_pose_lm_pnp(gt_Tcws, query_X_w, rand_R, scene_center, query_K,
pnp_x_2d, repro_thres):
N, _, H, W = query_X_w.shape
# recover original scene coordinates
query_X_3d_w = query_X_w.permute(0, 2, 3, 1).view(N, -1, 3)
rand_R_t = torch.transpose(rand_R, 1, 2).to(query_X_3d_w.device)
query_X_3d_w = batched_transpose(rand_R_t,
torch.zeros(N, 3).to(query_X_3d_w.device),
query_X_3d_w)
query_X_3d_w += scene_center.view(N, 1, 3)
query_X_3d_w = recover_original_scene_coordinates(query_X_w, rand_R,
scene_center)
query_X_3d_w = query_X_3d_w.view(N, H, W,
3).squeeze(0).detach().cpu().numpy()
# run Ransac PnP
lm_pnp_pose_vec, inlier_map = lm_pnp.compute_lm_pnp(
pnp_x_2d, query_X_3d_w, query_K, repro_thres, 128, 100)
R_res, _ = cv2.Rodrigues(lm_pnp_pose_vec[:3])
lm_pnp_pose = np.eye(4, dtype=np.float32)
lm_pnp_pose[:3, :3] = R_res
lm_pnp_pose[:3, 3] = lm_pnp_pose_vec[3:].ravel()
# measure accuracy
gt_pose = gt_Tcws.squeeze(0).detach().cpu().numpy()
R_acc = rel_rot_angle(lm_pnp_pose, gt_pose)
t_acc = rel_distance(lm_pnp_pose, gt_pose)
# ransc_inlier = None
return R_acc, t_acc, lm_pnp_pose, inlier_map
def verify_features(self, I_a, d_a, K, I_b, se3_gt, x, y, title):
"""
Extract feature pyramids f_a, f_b of I_a and I_b
Wrap f_b to f_a
Compute distances of a pixel in f_a with the neighbors of its corresponding pixels in f_b
:param I_a: Image of frame A, dim: (N, C, H, W)
:param d_a: Depth of frame A, dim: (N, 1, H, W)
:param K: intrinsic matrix at level 0: dim: (N, 3, 3)
:param I_b: Image of frame B, dim: (N, C, H, W)
:param se3_gt: Groundtruth of se3, dim: (N, 6)
:return:
"""
import banet_track.ba_debug as debug
(N, C, H, W) = I_a.shape
I_a.requires_grad_()
I_b.requires_grad_()
# Concate I_a and I_b
I = torch.cat([I_a, I_b], dim=0)
# Aggregate pyramid features
aggr_pyramid = self.aggregate_pyramid_features(self.backbone_net.forward(I))
aggr_pyramid_f_a = [f[:N, :, :, :] for f in aggr_pyramid]
aggr_pyramid_f_b = [f[N:, :, :, :] for f in aggr_pyramid]
for level in [2, 1, 0]:
(level_H, level_W) = self.level_dim_hw[level]
# Resize and Rescale the depth and the intrinsic matrix
rescale_ratio = 1.0 / math.pow(2, level)
level_K = rescale_ratio * K.detach() # dim: (N, 3, 3)
level_d_a = F.interpolate(d_a, scale_factor=rescale_ratio).detach() # dim: (N, 1, H, W)
# Cache several variables:
R, t = se3_exp(se3_gt)
x_a_2d = self.x_valid_2d[level] # dim: (N, H*W, 2)
X_a_3d = batched_pi_inv(level_K, x_a_2d,
level_d_a.view((N, level_H * level_W, 1)))
X_b_3d = batched_transpose(R, t, X_a_3d)
x_b_2d, _ = batched_pi(level_K, X_b_3d)
x_b_2d = module.batched_x_2d_normalize(float(level_H), float(level_W), x_b_2d).view(N, level_H, level_W, 2) # (N, H, W, 2)
# Wrap the feature
level_aggr_pyramid_f_b_wrap = batched_interp2d(aggr_pyramid_f_b[level], x_b_2d)
level_x = int(x * rescale_ratio)
level_y = int(y * rescale_ratio)
left = level_x - debug.similar_window_offset
left = left if left >= 0 else 0
right = level_x + debug.similar_window_offset
up = level_y - debug.similar_window_offset
up = up if up >= 0 else 0
down = level_y + debug.similar_window_offset
batch_distance = torch.norm(aggr_pyramid_f_a[level][:, :, up:down, left:right] - # (N, level_H, level_W)
level_aggr_pyramid_f_b_wrap[:, :, level_y:level_y+1, level_x:level_x+1], 2, 1)
show_multiple_img([{'img': I_a[0].detach().cpu().numpy().transpose(1, 2, 0), 'title': 'I_a'},
{'img': I_b[0].detach().cpu().numpy().transpose(1, 2, 0), 'title': 'I_b'},
{'img': batch_distance[0].detach().cpu().numpy(), 'title': 'feature distance', 'cmap':'gray'}],
title=title, num_cols=3)
def compute_pose_pnp_from_valid_pixels(gt_Tcws, query_X_w, rand_R,
scene_center, query_K, valid_pix_idx,
pnp_x_2d, repro_thres):
N, _, H, W = query_X_w.shape
# recover original scene coordinates
query_X_3d_w = query_X_w.permute(0, 2, 3, 1).view(N, -1, 3)
rand_R_t = torch.transpose(rand_R, 1, 2).to(query_X_3d_w.device)
query_X_3d_w = batched_transpose(rand_R_t,
torch.zeros(N, 3).to(query_X_3d_w.device),
query_X_3d_w)
query_X_3d_w += scene_center.view(N, 1, 3)
query_X_3d_w = recover_original_scene_coordinates(query_X_w, rand_R,
scene_center)
query_X_3d_w = query_X_3d_w.view(N, H, W,
3).squeeze(0).detach().cpu().numpy()
# select valid pixels with input index
x, y = valid_pix_idx
x_2d_valid = pnp_x_2d[y, x, :]
query_X_3d_valid = query_X_3d_w[y, x, :]
selected_pixels = query_X_3d_valid.shape[0]
query_X_3d_valid = query_X_3d_valid.reshape(1, selected_pixels, 3)
x_2d_valid = x_2d_valid.reshape(1, selected_pixels, 2)
# run Ransac PnP
dist = np.zeros(4)
k = query_K.squeeze(0).detach().cpu().numpy()
retval, R_res, t_res, ransc_inlier = cv2.solvePnPRansac(
query_X_3d_valid,
x_2d_valid,
k,
dist,
reprojectionError=repro_thres,
)
# print(retval)
# _, R_res, t_res = cv2.solvePnP(query_X_3d_valid, x_2d_valid, k, dist)#, flags=cv2.SOLVEPNP_EPNP)
R_res, _ = cv2.Rodrigues(R_res)
pnp_pose = np.eye(4, dtype=np.float32)
pnp_pose[:3, :3] = R_res
pnp_pose[:3, 3] = t_res.ravel()
# measure accuracy
gt_pose = gt_Tcws.squeeze(0).detach().cpu().numpy()
R_acc = rel_rot_angle(pnp_pose, gt_pose)
t_acc = rel_distance(pnp_pose, gt_pose)
# ransc_inlier = None
return R_acc, t_acc, pnp_pose, ransc_inlier
def recover_original_scene_coordinates(query_X_w, rand_R, scene_center):
N, _, H, W = query_X_w.shape
# recover original scene coordinates
query_X_3d_w = query_X_w.permute(0, 2, 3, 1).view(N, -1, 3)
rand_R_t = torch.transpose(rand_R, 1, 2).to(query_X_3d_w.device)
query_X_3d_w = batched_transpose(rand_R_t,
torch.zeros(N, 3).to(query_X_3d_w.device),
query_X_3d_w)
query_X_3d_w += scene_center.view(N, 1, 3)
query_X_3d_w = query_X_3d_w.view(N, H, W, 3)
return query_X_3d_w
def preprocess(sample_dict, pre_x2d, out_dim, rescale_dist=0.0):
rand_angle = np.random.random_sample() * 2.0 * np.pi
rand_R = quaternion_matrix(quaternion_about_axis(rand_angle, (0.0, 1.0, 0.0)))[:3, :3]
rand_R = torch.FloatTensor(rand_R).unsqueeze(0)
scene_rgb = sample_dict['frames_img'][:, :5, ...].cuda()
scene_depth = sample_dict['frames_depth'][:, :5, ...].cuda()
scene_K = sample_dict['frames_K'][:, :5, ...].cuda()
scene_Tcw = sample_dict['frames_Tcw'][:, :5, ...]
scene_ori_rgb = sample_dict['frames_ori_img'][:, :5, ...].cuda()
scene_neg_tags = sample_dict['frames_neg_tags'][:, :5, ...].cuda()
N, L, C, H, W = scene_rgb.shape
# scene_rgb = scene_rgb.view(N, L, C, H, W)
scene_depth = scene_depth.view(N * L, 1, H, W)
scene_K = scene_K.view(N * L, 3, 3)
scene_Tcw = scene_Tcw.view(N * L, 3, 4)
# generate 3D world position of scene
d = scene_depth.view(N * L, H * W, 1) # dim (N*L, H*W, 1)
X_3d = batched_pi_inv(scene_K, pre_x2d, d) # dim (N*L, H*W, 3)
Rwc, twc = batched_inv_pose(R=scene_Tcw[:, :3, :3],
t=scene_Tcw[:, :3, 3].squeeze(-1)) # dim (N*L, 3, 3), （N, 3)
X_world = batched_transpose(Rwc.cuda(), twc.cuda(), X_3d) # dim (N*L, H*W, 3)
X_world = X_world.contiguous().view(N, L * H * W, 3) # dim (N, L*H*W, 3)
scene_center = torch.mean(X_world, dim=1) # dim (N, 3)
X_world -= scene_center.view(N, 1, 3)
X_world = batched_transpose(rand_R.cuda().expand(N, 3, 3),
torch.zeros(1, 3, 1).cuda().expand(N, 3, 1),
X_world) # dim (N, L*H*W, 3), data augmentation
X_world = X_world.view(N, L, H, W, 3).permute(0, 1, 4, 2, 3).contiguous() # dim (N, L, 3, H, W)
# query image:
query_img = sample_dict['img']
query_ori_img = sample_dict['ori_img']
# compute multiscale ground truth query_X_worlds & valid_masks
query_X_worlds = []
valid_masks = []
out_H, out_W = out_dim
query_depth = sample_dict['depth'].cuda()
ori_query_depth = query_depth.clone()
N, C, H, W = query_depth.shape
for i in range(4):
query_depth_patch = F.unfold(
query_depth,
kernel_size=(H // out_H, W // out_W),
stride=(H // out_H, W // out_W)
).view(N, -1, out_H, out_W)
mask = torch.gt(query_depth_patch, 1e-5)
count = torch.sum(mask.float(), dim=1)
query_depth_down = torch.sum(query_depth_patch * mask.float(), dim=1) / \
torch.where(torch.le(count, 1e-5),
torch.full(count.shape, 1e6).to(count.device),
count) # (N, 1, out_H, out_W)
query_Tcw = sample_dict['Tcw']
query_K = sample_dict['K'].clone().cuda()
query_K[:, 0, 0] *= out_W / W
query_K[:, 0, 2] *= out_W / W
query_K[:, 1, 1] *= out_H / H
query_K[:, 1, 2] *= out_H / H
query_d = query_depth_down.view(N, out_H * out_W, 1) # dim (N, H*W, 1)
out_x_2d = x_2d_coords_torch(N, out_H, out_W).cuda().view(N, -1, 2)
query_X_3d = batched_pi_inv(query_K, out_x_2d, query_d) # dim (N, H*W, 3)
query_Rwc, query_twc = batched_inv_pose(R=query_Tcw[:, :3, :3],
t=query_Tcw[:, :3, 3].squeeze(-1)) # dim (N, 3, 3), （N, 3)
query_X_world = batched_transpose(query_Rwc.cuda(), query_twc.cuda(), query_X_3d) # dim (N, H*W, 3)
query_X_world -= scene_center.view(N, 1, 3)
query_X_world = batched_transpose(rand_R.cuda().expand(N, 3, 3),
torch.zeros(1, 3, 1).cuda().expand(N, 3, 1),
query_X_world) # dim (N, H*W, 3), data augmentation
query_X_world = query_X_world.permute(0, 2, 1).view(N, 3, out_H, out_W).contiguous() # dim (N, 3, H, W)
query_X_worlds.append(query_X_world.cuda())
valid_masks.append(torch.gt(query_depth_down, 1e-5).cuda().view(N, out_H, out_W))
if i == 3:
query_X_worlds.append(query_X_world.cuda())
valid_masks.append(torch.gt(query_depth_down, 1e-5).cuda().view(N, out_H, out_W))
out_H //= 2
out_W //= 2
# compute norm_query_Tcw for normalized scene coordinate
query_twc = query_twc.cuda() - scene_center.view(N, 3, 1)
norm_query_Twc = torch.cat([query_Rwc.cuda(), query_twc], dim=-1) # dim (N, 3, 4)
norm_query_Twc = torch.bmm(rand_R.cuda().expand(N, 3, 3), norm_query_Twc) # dim (N, 3, 4)
query_Rcw, query_tcw = batched_inv_pose(R=norm_query_Twc[:, :3, :3],
t=norm_query_Twc[:, :3, 3].squeeze(-1)) # dim (N, 3, 3), （N, 3)
norm_query_Tcw = torch.cat([query_Rcw, query_tcw.view(N, 3, 1)], dim=-1) # dim (N, 3, 4)
# compute down sampled query K
out_H, out_W = out_dim
query_K = sample_dict['K'].clone().cuda()
query_K[:, 0, 0] *= out_W / W
query_K[:, 0, 2] *= out_W / W
query_K[:, 1, 1] *= out_H / H
query_K[:, 1, 2] *= out_H / H
if rescale_dist > 0:
query_X_worlds, X_world, rescale_factor = rescale_scene_coords(query_X_worlds, X_world, scene_neg_tags, rescale_dist)
else:
rescale_factor = torch.ones(N)
scene_input = torch.cat((scene_rgb, X_world), dim=2)
return scene_input.cuda(), query_img.cuda(), query_X_worlds[::-1], valid_masks[::-1], \
scene_ori_rgb.cuda(), query_ori_img.cuda(), X_world.cuda(), \
torch.gt(scene_depth, 1e-5).cuda().view(N, L, H, W), norm_query_Tcw, query_K, scene_neg_tags, rescale_factor.cuda()
def valid(self, I_a, d_a, sel_a_indices, K, I_b, se3_gt, epoch):
"""
Pre cache the variable for prediction
:param I_a: Image of frame A, dim: (N, C, H, W)
:param d_a: Depth of frame A, dim: (N, 1, H, W)
:param sel_a_indices: (N, 3, M)
:param K: intrinsic matrix at level 0: dim: (N, 3, 3)
:param I_b: Image of frame B, dim: (N, C, H, W)
:param se3_gt: ground truth Pose
"""
(N, C, H, W) = I_a.shape
I_a.detach()
I_b.detach()
# Ground-truth pose
R_gt, t_gt = se3_exp(se3_gt)
# Concate I_a and I_b
I = torch.cat([I_a, I_b], dim=0)
# Aggregate pyramid features
aggr_pyramid = self.aggregate_pyramid_features(
self.backbone_net.forward(I))
aggr_pyramid_f_a = [f[:N, :, :, :] for f in aggr_pyramid]
aggr_pyramid_f_b = [f[N:, :, :, :] for f in aggr_pyramid]
# Init a se(3) vector and mark requires_grad = True
# alpha = torch.tensor([1e-4, 1e-4, 1e-4, 0.0, 0.0, 0.0]).repeat(N).view((N, 6)) # dim: (N, 6)
# factor = 0.3
# alpha = module.gen_random_alpha(se3_gt, rot_angle_rfactor=1.25, trans_vec_rfactor=0.16).view((N, 6)).cuda()
# alpha.requires_grad_()
T = torch.eye(4).view(1, 4, 4).repeat(N, 1, 1).detach()
init_T = T
pred_SE3_list = [
] # (num_level: low_res to high_res, num_iter_per_level)
gt_f_pair_list = []
lambda_weight = []
flow_list = []
for level in [2, 1, 0]:
pred_SE3_list.append([])
lambda_weight.append([])
flow_list.append([])
(level_H, level_W) = self.level_dim_hw[level]
M = sel_a_indices.shape[2] # number of selected pts
# Features on current level
f_a = aggr_pyramid_f_a[level]
f_b = aggr_pyramid_f_b[level]
f_b_grad = batched_gradient(f_b) # dim: (N, 2*C, H, W)
# Resize and Rescale the depth and the intrinsic matrix
rescale_ratio = 1.0 / math.pow(2, level)
level_K = rescale_ratio * K.detach() # dim: (N, 3, 3)
level_d_a = F.interpolate(
d_a, scale_factor=rescale_ratio).detach() # dim: (N, 1, H, W)
sel_a_idx = sel_a_indices[:,
level, :].view(N,
M).detach() # dim: (N, M)
# Cache several variables:
x_a_2d = self.x_train_2d[level] # dim: (N, H*W, 2)
X_a_3d = batched_pi_inv(level_K, x_a_2d,
level_d_a.view((N, level_H * level_W, 1)))
X_a_3d_sel = batched_index_select(X_a_3d, 1,
sel_a_idx) # dim: (N, M, 3)
""" Ground-truth correspondence for Regularizer
"""
f_C = f_a.shape[1]
X_b_3d_gt = batched_transpose(R_gt, t_gt, X_a_3d)
x_b_2d_gt, _ = batched_pi(level_K, X_b_3d_gt)
x_b_2d_gt = module.batched_x_2d_normalize(float(level_H),
float(level_W),
x_b_2d_gt).view(
N, level_H, level_W,
2) # (N, H, W, 2)
gt_f_wrap_b = batched_interp2d(f_b, x_b_2d_gt)
f_a_select = batched_index_select(
f_a.view(N, f_C, level_H * level_W), 2, sel_a_idx)
gt_f_wrap_b_select = batched_index_select(
gt_f_wrap_b.view(N, f_C, level_H * level_W), 2, sel_a_idx)
gt_f_pair_list.append((f_a_select, gt_f_wrap_b_select))
# Run iteration 3 times
for itr in range(0, 6):
T, r, delta_norm, lamb, flow = module.dm_levenberg_marquardt_itr(
T, X_a_3d, X_a_3d_sel, f_a, sel_a_idx, level_K, f_b,
f_b_grad, self.lambda_prediction, level)
pred_SE3_list[-1].append(T)
flow_list[-1].append((flow, x_b_2d_gt.detach()))
return pred_SE3_list, gt_f_pair_list, init_T.detach(), flow_list
def compute_pose(I_a,
d_a,
sel_a_idx,
I_b,
K,
alpha,
T_gt,
opt_max_itr=100,
opt_eps=1e-5):
# Debug assert
assert sel_a_indices.dtype == torch.int64
assert I_a.dtype == torch.float32
assert I_b.dtype == torch.float32
assert d_a.dtype == torch.float32
# Dimension
N, C, H, W = I_a.shape
# Pre-processing
# sel_a_idx = select_gradient_pixels(I_a, d_a, threshold=50.0)[: 2000]
M = sel_a_idx.shape[1]
I_b_grad = batched_gradient(
I_b
) # dim: (N, 2*C, H, W), (N, 0:C, H, W) = dI/dx, (N, C:2C, H, W) = dI/dy
assert H == d_a.shape[2]
assert W == d_a.shape[3]
# se(3) vector init
lambda_w = 0.2 * torch.ones(N, 6)
d_a = d_a.view((N, H * W, 1))
# Points' 3D Position at Frame a
x_a_2d = x_2d_coords_torch(N, H, W).view(N, H * W, 2)
X_a_3d = batched_pi_inv(K, x_a_2d, d_a)
X_a_3d_sel = batched_index_select(X_a_3d, 1, sel_a_idx)
# groundtruth wrap
alpha_gt = torch.tensor([0.5, 0.5, 0.5, 0.0, 0.0, 0.0]).repeat(N).view(
(N, 6))
R_gt, _ = se3_exp(alpha_gt)
I = torch.eye(3).view(1, 3, 3).expand(N, 3, 3).cuda()
zeros = torch.zeros_like(T_gt[:, :, 3]).cuda()
random_t = torch.zeros_like(T_gt[:, :, 3]).normal_(std=0.001)
#print('random_t:', random_t)
X_b_3d_gt = batched_transpose(R_gt, zeros, X_a_3d_sel)
x_b_2d_gt, _ = batched_pi(K, X_b_3d_gt)
for itr in range(0, opt_max_itr):
R, t = se3_exp(alpha)
X_b_3d = batched_transpose(R, t, X_a_3d_sel)
x_b_2d, _ = batched_pi(K, X_b_3d)
# Residual error
e = (x_b_2d_gt - x_b_2d).view(N, M * 2) # (N, H*W*2)
# Compute Jacobin Mat.
# Jacobi of Camera Pose: delta_u / delta_alpha
J = -J_camera_pose(X_a_3d_sel, K).view(N, M * 2, 6) # (N*M, 2, 6)
# x_b_2d = batched_x_2d_normalize(H, W, x_b_2d).view(N, H, W, 2) # (N, H, W, 2)
#
# # Wrap the image
# I_b_wrap = batched_interp2d(I_b, x_b_2d)
#
# # Residual error
# e = (I_a - I_b_wrap).view(N, C, H*W) # (N, C, H, W)
# e = batched_index_select(e, 2, sel_a_idx) # (N, C, M)
# e = e.transpose(1, 2).contiguous().view(N, M*C) # (N, M, C)
#
# # Compute Jacobin Mat.
# # Jacobi of Camera Pose: delta_u / delta_alpha
# du_d_alpha = J_camera_pose(X_a_3d_sel, K).view(N * M, 2, 6) # (N*M, 2, 6)
#
# # Jacobi of Image gradient: delta_I_b / delta_u
# dI_du = batched_interp2d(I_b_grad, x_b_2d) # (N, 2*C, H, W)
# dI_du = batched_index_select(dI_du.view(N, 2*C, H*W), 2, sel_a_idx) # (N, 2*C, M)
# dI_du = torch.transpose(dI_du, 1, 2).contiguous().view(N * M, 2, C) # (N*M, 2, C)
# dI_du = torch.transpose(dI_du, 1, 2) # (N*M, C, 2)
#
# # J = -dI_b/du * du/d_alpha
# J = -torch.bmm(dI_du, du_d_alpha).view(N, C*M, 6)
# Compute the update parameters
delta, delta_norm = gauss_newtown_update(J, e) # (N, 6), (N, 1)
max_norm = torch.max(delta_norm).item()
if max_norm < opt_eps:
print('break')
break
r_norm = torch.sum(e * e, dim=1) / M #2.0
print('Itr:', itr, 'r_norm=', torch.sqrt(r_norm), "update_norm=",
max_norm)
# Update the delta
alpha = alpha + delta
return R, t
K = torch.from_numpy(K_set).cuda().view((N, 3, 3)) T_gt = torch.from_numpy(T_gt_set).cuda().view((N, 3, 4)) alpha = torch.tensor([1e-4, 1e-4, 1e-4, 0.0, 0.0, 0.0]).repeat(N).view((N, 6)) R_, t_ = compute_pose(I_a, d_a, sel_a_indices, I_b, K, alpha, T_gt) print(R_, t_) print(T_gt[:, :, :3], T_gt[:, :, 3]) d_a = d_a.view((N, H * W, 1)) # Points' 3D Position at Frame a x_a_2d = x_2d_coords_torch(N, H, W).view(N, H * W, 2) X_a_3d = batched_pi_inv(K, x_a_2d, d_a) # groundtruth wrap X_b_3d_gt = batched_transpose(T_gt[:, :, :3], T_gt[:, :, 3], X_a_3d) x_b_2d_gt, _ = batched_pi(K, X_b_3d_gt) # # """ Test on 235 to 237 # """ # frame = frames.frames[128] # # Current frame # file_name = frame['file_name'] # img = cv2.imread(os.path.join(img_dir, file_name + '.jpg')).astype(np.float32) # (H, W, C) = img.shape # img = img.transpose((2, 0, 1)) / 255.0 # depth = load_depth_from_pgm(os.path.join(depth_dir, file_name + '.pgm')) # depth[depth < 1e-5] = 1e-5 # pose = frame['extrinsic_Tcw'] # K = K_from_frame(frame)
def preprocess_query(query_img,
query_depth,
query_ori_img,
query_Tcw,
ori_query_K,
scene_center,
rand_R,
out_dim=(48, 64)):
# compute multiscale ground truth query_X_worlds & valid_masks
query_X_worlds = []
valid_masks = []
out_H, out_W = out_dim
ori_query_depth = query_depth.clone()
N, C, H, W = query_depth.shape
for i in range(5):
query_depth_patch = F.unfold(query_depth,
kernel_size=(H // out_H, W // out_W),
stride=(H // out_H, W // out_W)).view(
N, -1, out_H, out_W)
mask = torch.gt(query_depth_patch, 1e-5)
count = torch.sum(mask.float(), dim=1)
query_depth_down = torch.sum(query_depth_patch * mask.float(), dim=1) /\
torch.where(torch.le(count, 1e-5),
torch.full(count.shape, 1e6).cuda(),
count) # (N, 1, out_H, out_W)
query_K = ori_query_K.clone().cuda()
query_K[:, 0, 0] *= out_W / W
query_K[:, 0, 2] *= out_W / W
query_K[:, 1, 1] *= out_H / H
query_K[:, 1, 2] *= out_H / H
query_d = query_depth_down.view(N, out_H * out_W, 1) # dim (N, H*W, 1)
out_x_2d = x_2d_coords_torch(N, out_H, out_W).cuda().view(N, -1, 2)
query_X_3d = batched_pi_inv(query_K, out_x_2d,
query_d) # dim (N, H*W, 3)
query_Rwc, query_twc = batched_inv_pose(
R=query_Tcw[:, :3, :3],
t=query_Tcw[:, :3, 3].squeeze(-1)) # dim (N, 3, 3), （N, 3)
query_X_world = batched_transpose(query_Rwc.cuda(), query_twc.cuda(),
query_X_3d) # dim (N, H*W, 3)
query_X_world -= scene_center.view(N, 1, 3)
query_X_world = batched_transpose(
rand_R.cuda().expand(N, 3, 3),
torch.zeros(1, 3, 1).cuda().expand(N, 3, 1),
query_X_world) # dim (N, H*W, 3), data augmentation
query_X_world = query_X_world.permute(0, 2, 1).view(
N, 3, out_H, out_W).contiguous() # dim (N, 3, H, W)
query_X_worlds.append(query_X_world.cuda())
valid_masks.append(
torch.gt(query_depth_down, 1e-5).cuda().view(N, out_H, out_W))
# if i == 3:
# query_X_worlds.append(query_X_world.cuda())
# valid_masks.append(torch.gt(query_depth_down, 1e-5).cuda().view(N, out_H, out_W))
out_H //= 2
out_W //= 2
# compute norm_query_Tcw for normalized scene coordinate
query_twc = query_twc.cuda() - scene_center.view(N, 3, 1)
norm_query_Twc = torch.cat([query_Rwc.cuda(), query_twc],
dim=-1) # dim (N, 3, 4)
norm_query_Twc = torch.bmm(rand_R.cuda().expand(N, 3, 3),
norm_query_Twc) # dim (N, 3, 4)
query_Rcw, query_tcw = batched_inv_pose(
R=norm_query_Twc[:, :3, :3],
t=norm_query_Twc[:, :3, 3].squeeze(-1)) # dim (N, 3, 3), （N, 3)
norm_query_Tcw = torch.cat([query_Rcw, query_tcw.view(N, 3, 1)],
dim=-1) # dim (N, 3, 4)
# compute down sampled query K
out_H, out_W = out_dim
query_K = ori_query_K.clone().cuda()
query_K[:, 0, 0] *= out_W / W
query_K[:, 0, 2] *= out_W / W
query_K[:, 1, 1] *= out_H / H
query_K[:, 1, 2] *= out_H / H
return query_img.cuda(), query_X_worlds[::-1], valid_masks[::-1], query_ori_img.cuda(), \
scene_center.cuda(), query_Tcw.cuda(), query_K.cuda(), rand_R.expand(N, 3, 3).cuda()
