Delete process.py

process.py

@@ -1,126 +0,0 @@
-import numpy as np
-import torch
-import os
-import re
-from packaging import version as pver
-
-def custom_meshgrid(*args):
-    # ref: https://pytorch.org/docs/stable/generated/torch.meshgrid.html?highlight=meshgrid#torch.meshgrid
-    if pver.parse(torch.__version__) < pver.parse('1.10'):
-        return torch.meshgrid(*args)
-    else:
-        return torch.meshgrid(*args, indexing='ij')
-
-
-def ray_condition(K, c2w, H, W, device, flip_flag=None):
-    # c2w: B, V, 4, 4
-    # K: B, V, 4
-
-    B, V = K.shape[:2]
-
-    j, i = custom_meshgrid(
-        torch.linspace(0, H - 1, H, device=device, dtype=c2w.dtype),
-        torch.linspace(0, W - 1, W, device=device, dtype=c2w.dtype),
-    )
-    i = i.reshape([1, 1, H * W]).expand([B, V, H * W]) + 0.5          # [B, V, HxW]
-    j = j.reshape([1, 1, H * W]).expand([B, V, H * W]) + 0.5          # [B, V, HxW]
-
-    n_flip = torch.sum(flip_flag).item() if flip_flag is not None else 0
-    if n_flip > 0:
-        j_flip, i_flip = custom_meshgrid(
-            torch.linspace(0, H - 1, H, device=device, dtype=c2w.dtype),
-            torch.linspace(W - 1, 0, W, device=device, dtype=c2w.dtype)
-        )
-        i_flip = i_flip.reshape([1, 1, H * W]).expand(B, 1, H * W) + 0.5
-        j_flip = j_flip.reshape([1, 1, H * W]).expand(B, 1, H * W) + 0.5
-        i[:, flip_flag, ...] = i_flip
-        j[:, flip_flag, ...] = j_flip
-
-    fx, fy, cx, cy = K.chunk(4, dim=-1)     # B,V, 1
-
-    zs = torch.ones_like(i)                 # [B, V, HxW]
-    xs = (i - cx) / fx * zs
-    ys = (j - cy) / fy * zs
-    zs = zs.expand_as(ys)
-
-    directions = torch.stack((xs, ys, zs), dim=-1)              # B, V, HW, 3
-    directions = directions / directions.norm(dim=-1, keepdim=True)             # B, V, HW, 3
-
-    rays_d = directions @ c2w[..., :3, :3].transpose(-1, -2)        # B, V, HW, 3
-    rays_o = c2w[..., :3, 3]                                        # B, V, 3
-    rays_o = rays_o[:, :, None].expand_as(rays_d)                   # B, V, HW, 3
-    # c2w @ dirctions
-    rays_dxo = torch.cross(rays_o, rays_d)                          # B, V, HW, 3
-    plucker = torch.cat([rays_dxo, rays_d], dim=-1)
-    plucker = plucker.reshape(B, c2w.shape[1], H, W, 6)             # B, V, H, W, 6
-    # plucker = plucker.permute(0, 1, 4, 2, 3)
-    return plucker
-
-def plucker_from_camera_params(
-    K, R, t, W0, H0, cw_np,
-    H, W,
-    device='cpu',
-    normalize_intrinsics=True
-):
-    """
-    K, R, t, W0, H0 同 get_camera_params 返回，
-    cw_np 是 camera.matrix_world 转成的 numpy.ndarray (4×4)。
-    H, W 是你要的光线网格尺寸。
-    """
-    # 1) 归一化 intrinsics
-    if normalize_intrinsics:
-        fx = K[0,0] / W0
-        fy = K[1,1] / H0
-        cx = K[0,2] / W0
-        cy = K[1,2] / H0
-    else:
-        fx, fy = K[0,0], K[1,1]
-        cx, cy = K[0,2], K[1,2]
-
-    intrinsics = torch.tensor(
-        [[[fx, fy, cx, cy]]],
-        dtype=torch.float32,
-        device=device
-    )  # [1,1,4]
-
-    # 2) cw_np 就是 c2w
-    #    直接转成张量，不要再调用 Matrix 方法
-    c2w = torch.tensor(cw_np, dtype=torch.float32, device=device)
-    c2w = c2w.unsqueeze(0).unsqueeze(0)  # [1,1,4,4]
-
-    # 3) 调用 ray_condition
-    flip_flag = torch.zeros(1, dtype=torch.bool, device=device)
-    plucker = ray_condition(
-        intrinsics,
-        c2w,
-        H, W,
-        device=device,
-        flip_flag=flip_flag
-    )  # [1,1,H,W,6]
-
-    return plucker[0,0]  # [6, H, W]
-
-
-dir_path = "1_19_Idle"
-dict_files = sorted([os.path.join(dir_path, file) for file in os.listdir(dir_path) if file.endswith(".npy") and "dict" in file])
-
-
-for dict_file in dict_files:
-    params = np.load(dict_file, allow_pickle=True).item()
-    match_view = re.search(r'view(\d+)', str(dict_file))
-    if match_view:
-        view = int(match_view.group(1))
-    else:
-        view = None
-
-    plucker_embedding = plucker_from_camera_params(
-        params['K'], params['R'], params['t'], params['W0'], params['H0'], params['cw'],
-        480, 720,
-        device='cpu',
-        normalize_intrinsics=True
-    )
-    plucker_embedding = torch.tensor(plucker_embedding).permute(2,0,1)
-    o = params['cw'][:3, 3] 
-    r = np.linalg.norm(o)
-    plucker_embedding[:3] = plucker_embedding[:3] / r  
-    torch.save(plucker_embedding, f"{view}.pt")