--- a
+++ b/utils/direct_field/utils_df.py
@@ -0,0 +1,34 @@
+import torch
+import numpy as np
+from scipy.ndimage import distance_transform_edt as distance
+from utils.utils_loss import one_hot, simplex, class2one_hot
+
+def one_hot2dist(seg: np.ndarray) -> np.ndarray:
+    assert one_hot(torch.Tensor(seg), axis=0)
+    C: int = len(seg)
+
+    res = np.zeros_like(seg)
+    for c in range(C):
+        posmask = seg[c].astype(np.bool)
+
+        if posmask.any():
+            negmask = ~posmask
+            res[c] = distance(negmask) * negmask - (distance(posmask) - 1) * posmask
+    return res
+
+def class2dist(seg: np.ndarray, C=4) -> np.ndarray:
+    """ res: (C, H, W)
+    """
+    if seg.ndim == 2:
+        seg_tensor = torch.Tensor(seg)
+    elif seg.ndim == 3:
+        seg_tensor = torch.Tensor(seg[0])
+    elif seg.ndim == 4:
+        seg_tensor = torch.Tensor(seg[0, ..., 0])
+
+    seg_onehot = class2one_hot(seg_tensor, C).to(torch.float32)
+
+    assert simplex(seg_onehot)
+    res = one_hot2dist(seg_onehot[0].numpy())
+    return res
+